{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Train Quartznet"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Paper related https://arxiv.org/pdf/1910.10261.pdf"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"\n",
"\n",
"This tutorial is available as an IPython notebook at [malaya-speech/example/train-asr](https://github.com/huseinzol05/malaya-speech/tree/master/example/train-asr).\n",
" \n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"\n",
"\n",
"This example trained on a very small dataset, do not use it for production.\n",
" \n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Model Interface\n",
"\n",
"To initiate Quartznet,\n",
"\n",
"```python\n",
"import malaya_speech.train.model.quartznet as quartznet\n",
"\n",
"x = tf.placeholder(tf.float32, [None, None, num_features])\n",
"x_len = tf.placeholder(tf.int32, [None])\n",
"quartznet.Model(x, x_len)\n",
"```\n",
"\n",
"Required input lengths, so the model able to calculate output lengths."
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import os\n",
"\n",
"os.environ['CUDA_VISIBLE_DEVICES'] = '0'"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"WARNING:tensorflow:From /home/husein/malaya-speech/malaya_speech/train/model/quartznet/layer.py:6: The name tf.layers.Conv1D is deprecated. Please use tf.compat.v1.layers.Conv1D instead.\n",
"\n",
"WARNING:tensorflow:\n",
"The TensorFlow contrib module will not be included in TensorFlow 2.0.\n",
"For more information, please see:\n",
" * https://github.com/tensorflow/community/blob/master/rfcs/20180907-contrib-sunset.md\n",
" * https://github.com/tensorflow/addons\n",
" * https://github.com/tensorflow/io (for I/O related ops)\n",
"If you depend on functionality not listed there, please file an issue.\n",
"\n"
]
}
],
"source": [
"import tensorflow as tf\n",
"from glob import glob\n",
"import malaya_speech.config\n",
"import malaya_speech.train.model.quartznet as quartznet\n",
"import malaya_speech.train.model.ctc as ctc\n",
"import malaya_speech.train as train\n",
"import numpy as np"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Load Featurizer\n",
"\n",
"We prefer use `malaya_speech.tf_featurization.STTFeaturizer` for ASR."
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [],
"source": [
"config = malaya_speech.config.transducer_featurizer_config\n",
"featurizer = malaya_speech.tf_featurization.STTFeaturizer(**config)\n",
"n_mels = config['num_feature_bins']\n",
"initial_learning_rate = 1e-3\n",
"max_gradient_norm = 5.0\n",
"train_steps = 1000"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Define Data Pipeline"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [],
"source": [
"def preprocess_inputs(example):\n",
" s = featurizer.vectorize(example['waveforms'])\n",
" mel_fbanks = tf.reshape(s, (-1, n_mels))\n",
" length = tf.cast(tf.shape(mel_fbanks)[0], tf.int32)\n",
" length = tf.expand_dims(length, 0)\n",
" example['inputs'] = mel_fbanks\n",
" example['inputs_length'] = length\n",
"\n",
" return example"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [],
"source": [
"def parse(serialized_example):\n",
"\n",
" data_fields = {\n",
" 'waveforms': tf.VarLenFeature(tf.float32),\n",
" 'targets': tf.VarLenFeature(tf.int64),\n",
" }\n",
" features = tf.parse_single_example(\n",
" serialized_example, features = data_fields\n",
" )\n",
" for k in features.keys():\n",
" features[k] = features[k].values\n",
"\n",
" features = preprocess_inputs(features)\n",
"\n",
" keys = list(features.keys())\n",
" for k in keys:\n",
" if k not in ['inputs', 'inputs_length', 'targets']:\n",
" features.pop(k, None)\n",
"\n",
" return features\n",
"\n",
"\n",
"def get_dataset(path, batch_size = 32, shuffle_size = 32, thread_count = 24):\n",
" def get():\n",
" files = glob(path)\n",
" dataset = tf.data.TFRecordDataset(files)\n",
" dataset = dataset.shuffle(shuffle_size)\n",
" dataset = dataset.map(parse, num_parallel_calls = thread_count)\n",
" dataset = dataset.padded_batch(\n",
" batch_size,\n",
" padded_shapes = {\n",
" 'inputs': tf.TensorShape([None, n_mels]),\n",
" 'inputs_length': tf.TensorShape([None]),\n",
" 'targets': tf.TensorShape([None]),\n",
" },\n",
" padding_values = {\n",
" 'inputs': tf.constant(0, dtype = tf.float32),\n",
" 'inputs_length': tf.constant(0, dtype = tf.int32),\n",
" 'targets': tf.constant(0, dtype = tf.int64),\n",
" },\n",
" )\n",
" dataset = dataset.repeat()\n",
" return dataset\n",
"\n",
" return get"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Define Model definition\n",
"\n",
"1. Use malaya_speech CTC loss,\n",
"\n",
"```python\n",
"mean_error, sum_error, sum_weight = ctc.loss.ctc_loss(\n",
" logits, targets_int32, seq_lens\n",
")\n",
"```\n",
"\n",
"2. Use malaya_speech CTC sequence accuracy,\n",
"\n",
"```python\n",
"accuracy = ctc.metrics.ctc_sequence_accuracy(\n",
" logits, targets_int32, seq_lens\n",
")\n",
"```\n",
"\n",
"This will automatically recorded in Tensorboard.\n",
"\n",
"3. Use malaya_speech metrics for Evaluation session,\n",
"\n",
"```python\n",
"elif mode == tf.estimator.ModeKeys.EVAL:\n",
"\n",
" estimator_spec = tf.estimator.EstimatorSpec(\n",
" mode = tf.estimator.ModeKeys.EVAL,\n",
" loss = loss,\n",
" eval_metric_ops = {\n",
" 'accuracy': ctc.metrics.ctc_sequence_accuracy_estimator(\n",
" logits, targets_int32, seq_lens\n",
" ),\n",
" 'WER': ctc.metrics.word_error_rate_estimator(\n",
" logits, targets_int32\n",
" ),\n",
" },\n",
" )\n",
"```\n",
"\n",
"If we passed directly tensor values to `eval_metric_ops`, Tensorflow will throw error, required update operations. So just stick with that implementation."
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [],
"source": [
"def model_fn(features, labels, mode, params):\n",
"\n",
" model = quartznet.Model(\n",
" features['inputs'], features['inputs_length'][:, 0], mode = 'train'\n",
" )\n",
" logits = tf.layers.dense(\n",
" model.logits['outputs'], malaya_speech.char.VOCAB_SIZE\n",
" )\n",
" seq_lens = model.logits['src_length']\n",
"\n",
" targets_int32 = tf.cast(features['targets'], tf.int32)\n",
"\n",
" mean_error, sum_error, sum_weight = ctc.loss.ctc_loss(\n",
" logits, targets_int32, seq_lens\n",
" )\n",
"\n",
" loss = mean_error\n",
" accuracy = ctc.metrics.ctc_sequence_accuracy(\n",
" logits, targets_int32, seq_lens\n",
" )\n",
"\n",
" tf.identity(loss, 'train_loss')\n",
" tf.identity(accuracy, name = 'train_accuracy')\n",
"\n",
" if mode == tf.estimator.ModeKeys.TRAIN:\n",
" global_step = tf.train.get_or_create_global_step()\n",
" learning_rate = tf.train.polynomial_decay(\n",
" initial_learning_rate,\n",
" global_step,\n",
" train_steps,\n",
" end_learning_rate = 1e-4,\n",
" power = 1.0,\n",
" cycle = False,\n",
" )\n",
" optimizer = tf.train.AdamOptimizer(learning_rate = learning_rate)\n",
" train_op = optimizer.minimize(loss, global_step = global_step)\n",
" estimator_spec = tf.estimator.EstimatorSpec(\n",
" mode = mode, loss = loss, train_op = train_op\n",
" )\n",
"\n",
" elif mode == tf.estimator.ModeKeys.EVAL:\n",
"\n",
" estimator_spec = tf.estimator.EstimatorSpec(\n",
" mode = tf.estimator.ModeKeys.EVAL,\n",
" loss = loss,\n",
" eval_metric_ops = {\n",
" 'accuracy': ctc.metrics.ctc_sequence_accuracy_estimator(\n",
" logits, targets_int32, seq_lens\n",
" ),\n",
" 'WER': ctc.metrics.word_error_rate_estimator(\n",
" logits, targets_int32\n",
" ),\n",
" },\n",
" )\n",
"\n",
" return estimator_spec"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [],
"source": [
"train_hooks = [\n",
" tf.train.LoggingTensorHook(\n",
" ['train_accuracy', 'train_loss'], every_n_iter = 1\n",
" )\n",
"]\n",
"train_dataset = get_dataset('tolong-sebut/data/tolong-sebut-train*')\n",
"dev_dataset = get_dataset('tolong-sebut/data/tolong-sebut-dev*')"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [],
"source": [
"!rm -rf asr-quartznet"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Initate training session"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"WARNING:tensorflow:From /home/husein/malaya-speech/malaya_speech/train/__init__.py:38: The name tf.logging.set_verbosity is deprecated. Please use tf.compat.v1.logging.set_verbosity instead.\n",
"\n",
"WARNING:tensorflow:From /home/husein/malaya-speech/malaya_speech/train/__init__.py:38: The name tf.logging.INFO is deprecated. Please use tf.compat.v1.logging.INFO instead.\n",
"\n",
"INFO:tensorflow:Using config: {'_model_dir': 'asr-quartznet', '_tf_random_seed': None, '_save_summary_steps': 100, '_save_checkpoints_steps': 200, '_save_checkpoints_secs': None, '_session_config': allow_soft_placement: true\n",
"graph_options {\n",
" rewrite_options {\n",
" meta_optimizer_iterations: ONE\n",
" }\n",
"}\n",
", '_keep_checkpoint_max': 5, '_keep_checkpoint_every_n_hours': 10000, '_log_step_count_steps': 1, '_train_distribute': None, '_device_fn': None, '_protocol': None, '_eval_distribute': None, '_experimental_distribute': None, '_experimental_max_worker_delay_secs': None, '_session_creation_timeout_secs': 7200, '_service': None, '_cluster_spec': , '_task_type': 'worker', '_task_id': 0, '_global_id_in_cluster': 0, '_master': '', '_evaluation_master': '', '_is_chief': True, '_num_ps_replicas': 0, '_num_worker_replicas': 1}\n",
"INFO:tensorflow:Not using Distribute Coordinator.\n",
"INFO:tensorflow:Running training and evaluation locally (non-distributed).\n",
"INFO:tensorflow:Start train and evaluate loop. The evaluate will happen after every checkpoint. Checkpoint frequency is determined based on RunConfig arguments: save_checkpoints_steps 200 or save_checkpoints_secs None.\n",
"WARNING:tensorflow:From /home/husein/.local/lib/python3.6/site-packages/tensorflow_core/python/training/training_util.py:236: Variable.initialized_value (from tensorflow.python.ops.variables) is deprecated and will be removed in a future version.\n",
"Instructions for updating:\n",
"Use Variable.read_value. Variables in 2.X are initialized automatically both in eager and graph (inside tf.defun) contexts.\n",
"WARNING:tensorflow:From /home/husein/.local/lib/python3.6/site-packages/tensorflow_core/python/autograph/converters/directives.py:119: The name tf.VarLenFeature is deprecated. Please use tf.io.VarLenFeature instead.\n",
"\n",
"WARNING:tensorflow:From /home/husein/.local/lib/python3.6/site-packages/tensorflow_core/python/autograph/converters/directives.py:119: The name tf.parse_single_example is deprecated. Please use tf.io.parse_single_example instead.\n",
"\n",
"INFO:tensorflow:Calling model_fn.\n",
"WARNING:tensorflow:From /home/husein/malaya-speech/malaya_speech/train/model/quartznet/abstract.py:141: The name tf.variable_scope is deprecated. Please use tf.compat.v1.variable_scope instead.\n",
"\n",
"WARNING:tensorflow:From /home/husein/malaya-speech/malaya_speech/train/model/quartznet/layer.py:327: separable_conv1d (from tensorflow.python.layers.convolutional) is deprecated and will be removed in a future version.\n",
"Instructions for updating:\n",
"Use `tf.keras.layers.SeparableConv1D` instead.\n",
"WARNING:tensorflow:From /home/husein/.local/lib/python3.6/site-packages/tensorflow_core/python/layers/convolutional.py:971: Layer.apply (from tensorflow.python.keras.engine.base_layer) is deprecated and will be removed in a future version.\n",
"Instructions for updating:\n",
"Please use `layer.__call__` method instead.\n",
"WARNING:tensorflow:From /home/husein/malaya-speech/malaya_speech/train/model/quartznet/layer.py:358: batch_normalization (from tensorflow.python.layers.normalization) is deprecated and will be removed in a future version.\n",
"Instructions for updating:\n",
"Use keras.layers.BatchNormalization instead. In particular, `tf.control_dependencies(tf.GraphKeys.UPDATE_OPS)` should not be used (consult the `tf.keras.layers.batch_normalization` documentation).\n",
"WARNING:tensorflow:From /home/husein/malaya-speech/malaya_speech/train/model/quartznet/abstract.py:369: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
"Instructions for updating:\n",
"Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
"WARNING:tensorflow:From /home/husein/malaya-speech/malaya_speech/train/model/quartznet/layer.py:340: conv1d (from tensorflow.python.layers.convolutional) is deprecated and will be removed in a future version.\n",
"Instructions for updating:\n",
"Use `tf.keras.layers.Conv1D` instead.\n",
"WARNING:tensorflow:From :7: dense (from tensorflow.python.layers.core) is deprecated and will be removed in a future version.\n",
"Instructions for updating:\n",
"Use keras.layers.Dense instead.\n",
"WARNING:tensorflow:From /home/husein/malaya-speech/malaya_speech/train/model/ctc/loss.py:14: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
"Instructions for updating:\n",
"Use `tf.cast` instead.\n",
"WARNING:tensorflow:From /home/husein/.local/lib/python3.6/site-packages/tensorflow_core/python/ops/array_ops.py:1475: where (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
"Instructions for updating:\n",
"Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
"WARNING:tensorflow:From /home/husein/malaya-speech/malaya_speech/train/model/ctc/loss.py:20: The name tf.nn.ctc_loss is deprecated. Please use tf.compat.v1.nn.ctc_loss instead.\n",
"\n",
"WARNING:tensorflow:From /home/husein/.local/lib/python3.6/site-packages/tensorflow_core/python/util/deprecation.py:507: calling count_nonzero (from tensorflow.python.ops.math_ops) with axis is deprecated and will be removed in a future version.\n",
"Instructions for updating:\n",
"reduction_indices is deprecated, use axis instead\n",
"INFO:tensorflow:Done calling model_fn.\n",
"INFO:tensorflow:Create CheckpointSaverHook.\n",
"INFO:tensorflow:Graph was finalized.\n",
"INFO:tensorflow:Running local_init_op.\n",
"INFO:tensorflow:Done running local_init_op.\n",
"INFO:tensorflow:Saving checkpoints for 0 into asr-quartznet/model.ckpt.\n",
"INFO:tensorflow:train_accuracy = 0.0077021825, train_loss = 710.26636\n",
"INFO:tensorflow:loss = 710.26636, step = 0\n",
"INFO:tensorflow:global_step/sec: 0.157221\n",
"INFO:tensorflow:train_accuracy = 0.029299363, train_loss = 529.1721 (6.361 sec)\n",
"INFO:tensorflow:loss = 529.1721, step = 1 (6.359 sec)\n",
"INFO:tensorflow:global_step/sec: 1.40909\n",
"INFO:tensorflow:train_accuracy = 0.014925373, train_loss = 184.64783 (0.709 sec)\n",
"INFO:tensorflow:loss = 184.64783, step = 2 (0.709 sec)\n",
"INFO:tensorflow:global_step/sec: 1.47474\n",
"INFO:tensorflow:train_accuracy = 0.0125470515, train_loss = 136.85947 (0.678 sec)\n",
"INFO:tensorflow:loss = 136.85947, step = 3 (0.678 sec)\n",
"INFO:tensorflow:global_step/sec: 1.68355\n",
"INFO:tensorflow:train_accuracy = 0.022427442, train_loss = 120.88869 (0.594 sec)\n",
"INFO:tensorflow:loss = 120.88869, step = 4 (0.594 sec)\n",
"INFO:tensorflow:global_step/sec: 1.40731\n",
"INFO:tensorflow:train_accuracy = 0.012919896, train_loss = 99.73471 (0.710 sec)\n",
"INFO:tensorflow:loss = 99.73471, step = 5 (0.710 sec)\n",
"INFO:tensorflow:global_step/sec: 1.36188\n",
"INFO:tensorflow:train_accuracy = 0.016774194, train_loss = 100.694885 (0.734 sec)\n",
"INFO:tensorflow:loss = 100.694885, step = 6 (0.734 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69408\n",
"INFO:tensorflow:train_accuracy = 0.015625, train_loss = 86.29912 (0.590 sec)\n",
"INFO:tensorflow:loss = 86.29912, step = 7 (0.592 sec)\n",
"INFO:tensorflow:global_step/sec: 1.3788\n",
"INFO:tensorflow:train_accuracy = 0.018181818, train_loss = 85.87342 (0.725 sec)\n",
"INFO:tensorflow:loss = 85.87342, step = 8 (0.724 sec)\n",
"INFO:tensorflow:global_step/sec: 1.60367\n",
"INFO:tensorflow:train_accuracy = 0.014285714, train_loss = 86.42359 (0.623 sec)\n",
"INFO:tensorflow:loss = 86.42359, step = 9 (0.623 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64712\n",
"INFO:tensorflow:train_accuracy = 0.01810585, train_loss = 70.12548 (0.607 sec)\n",
"INFO:tensorflow:loss = 70.12548, step = 10 (0.607 sec)\n",
"INFO:tensorflow:global_step/sec: 1.25871\n",
"INFO:tensorflow:train_accuracy = 0.012674271, train_loss = 84.80704 (0.795 sec)\n",
"INFO:tensorflow:loss = 84.80704, step = 11 (0.795 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75393\n",
"INFO:tensorflow:train_accuracy = 0.026041666, train_loss = 81.86655 (0.570 sec)\n",
"INFO:tensorflow:loss = 81.86655, step = 12 (0.570 sec)\n",
"INFO:tensorflow:global_step/sec: 1.59815\n",
"INFO:tensorflow:train_accuracy = 0.011795544, train_loss = 69.670364 (0.626 sec)\n",
"INFO:tensorflow:loss = 69.670364, step = 13 (0.625 sec)\n",
"INFO:tensorflow:global_step/sec: 1.59812\n",
"INFO:tensorflow:train_accuracy = 0.016927084, train_loss = 74.92718 (0.626 sec)\n",
"INFO:tensorflow:loss = 74.92718, step = 14 (0.626 sec)\n",
"INFO:tensorflow:global_step/sec: 1.54822\n",
"INFO:tensorflow:train_accuracy = 0.007989348, train_loss = 67.74375 (0.646 sec)\n",
"INFO:tensorflow:loss = 67.74375, step = 15 (0.647 sec)\n",
"INFO:tensorflow:global_step/sec: 1.57157\n",
"INFO:tensorflow:train_accuracy = 0.03192848, train_loss = 77.726746 (0.636 sec)\n",
"INFO:tensorflow:loss = 77.726746, step = 16 (0.636 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7516\n",
"INFO:tensorflow:train_accuracy = 0.031128405, train_loss = 77.6071 (0.571 sec)\n",
"INFO:tensorflow:loss = 77.6071, step = 17 (0.572 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69397\n",
"INFO:tensorflow:train_accuracy = 0.0078023407, train_loss = 68.87024 (0.590 sec)\n",
"INFO:tensorflow:loss = 68.87024, step = 18 (0.590 sec)\n",
"INFO:tensorflow:global_step/sec: 1.55234\n",
"INFO:tensorflow:train_accuracy = 0.011968086, train_loss = 68.68738 (0.644 sec)\n",
"INFO:tensorflow:loss = 68.68738, step = 19 (0.644 sec)\n",
"INFO:tensorflow:global_step/sec: 1.35571\n",
"INFO:tensorflow:train_accuracy = 0.017971758, train_loss = 74.069534 (0.737 sec)\n",
"INFO:tensorflow:loss = 74.069534, step = 20 (0.738 sec)\n",
"INFO:tensorflow:global_step/sec: 1.55871\n",
"INFO:tensorflow:train_accuracy = 0.011306533, train_loss = 69.41161 (0.642 sec)\n",
"INFO:tensorflow:loss = 69.41161, step = 21 (0.641 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7586\n",
"INFO:tensorflow:train_accuracy = 0.023684211, train_loss = 75.44877 (0.569 sec)\n",
"INFO:tensorflow:loss = 75.44877, step = 22 (0.569 sec)\n",
"INFO:tensorflow:global_step/sec: 1.67861\n",
"INFO:tensorflow:train_accuracy = 0.03846154, train_loss = 66.29564 (0.596 sec)\n",
"INFO:tensorflow:loss = 66.29564, step = 23 (0.596 sec)\n",
"INFO:tensorflow:global_step/sec: 1.57344\n",
"INFO:tensorflow:train_accuracy = 0.014285714, train_loss = 65.60063 (0.635 sec)\n",
"INFO:tensorflow:loss = 65.60063, step = 24 (0.635 sec)\n",
"INFO:tensorflow:global_step/sec: 1.55518\n",
"INFO:tensorflow:train_accuracy = 0.020547945, train_loss = 63.272343 (0.643 sec)\n",
"INFO:tensorflow:loss = 63.272343, step = 25 (0.643 sec)\n",
"INFO:tensorflow:global_step/sec: 1.67928\n",
"INFO:tensorflow:train_accuracy = 0.030144168, train_loss = 62.714897 (0.595 sec)\n",
"INFO:tensorflow:loss = 62.714897, step = 26 (0.596 sec)\n",
"INFO:tensorflow:global_step/sec: 1.31984\n",
"INFO:tensorflow:train_accuracy = 0.013613861, train_loss = 74.59552 (0.758 sec)\n",
"INFO:tensorflow:loss = 74.59552, step = 27 (0.758 sec)\n",
"INFO:tensorflow:global_step/sec: 1.96467\n",
"INFO:tensorflow:train_accuracy = 0.031094527, train_loss = 66.440704 (0.509 sec)\n",
"INFO:tensorflow:loss = 66.440704, step = 28 (0.508 sec)\n",
"INFO:tensorflow:global_step/sec: 1.67486\n",
"INFO:tensorflow:train_accuracy = 0.038873993, train_loss = 60.191017 (0.597 sec)\n",
"INFO:tensorflow:loss = 60.191017, step = 29 (0.597 sec)\n",
"INFO:tensorflow:global_step/sec: 1.3225\n",
"INFO:tensorflow:train_accuracy = 0.016209476, train_loss = 78.92613 (0.757 sec)\n",
"INFO:tensorflow:loss = 78.92613, step = 30 (0.757 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64088\n",
"INFO:tensorflow:train_accuracy = 0.04199475, train_loss = 65.53371 (0.609 sec)\n",
"INFO:tensorflow:loss = 65.53371, step = 31 (0.609 sec)\n",
"INFO:tensorflow:global_step/sec: 1.47649\n",
"INFO:tensorflow:train_accuracy = 0.0325, train_loss = 74.510284 (0.677 sec)\n",
"INFO:tensorflow:loss = 74.510284, step = 32 (0.677 sec)\n",
"INFO:tensorflow:global_step/sec: 1.47434\n",
"INFO:tensorflow:train_accuracy = 0.051157124, train_loss = 70.60351 (0.678 sec)\n",
"INFO:tensorflow:loss = 70.60351, step = 33 (0.678 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89935\n",
"INFO:tensorflow:train_accuracy = 0.02617801, train_loss = 67.7594 (0.526 sec)\n",
"INFO:tensorflow:loss = 67.7594, step = 34 (0.526 sec)\n",
"INFO:tensorflow:global_step/sec: 1.47234\n",
"INFO:tensorflow:train_accuracy = 0.025252525, train_loss = 69.94642 (0.679 sec)\n",
"INFO:tensorflow:loss = 69.94642, step = 35 (0.679 sec)\n",
"INFO:tensorflow:global_step/sec: 1.31145\n",
"INFO:tensorflow:train_accuracy = 0.018939395, train_loss = 70.5148 (0.763 sec)\n",
"INFO:tensorflow:loss = 70.5148, step = 36 (0.763 sec)\n",
"INFO:tensorflow:global_step/sec: 1.4745\n",
"INFO:tensorflow:train_accuracy = 0.01754386, train_loss = 68.67237 (0.678 sec)\n",
"INFO:tensorflow:loss = 68.67237, step = 37 (0.678 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88099\n",
"INFO:tensorflow:train_accuracy = 0.028423773, train_loss = 70.48649 (0.531 sec)\n",
"INFO:tensorflow:loss = 70.48649, step = 38 (0.531 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64177\n",
"INFO:tensorflow:train_accuracy = 0.035526317, train_loss = 67.249985 (0.610 sec)\n",
"INFO:tensorflow:loss = 67.249985, step = 39 (0.610 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71506\n",
"INFO:tensorflow:train_accuracy = 0.047425475, train_loss = 57.39673 (0.583 sec)\n",
"INFO:tensorflow:loss = 57.39673, step = 40 (0.583 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7518\n",
"INFO:tensorflow:train_accuracy = 0.03508772, train_loss = 66.17575 (0.571 sec)\n",
"INFO:tensorflow:loss = 66.17575, step = 41 (0.571 sec)\n",
"INFO:tensorflow:global_step/sec: 1.45753\n",
"INFO:tensorflow:train_accuracy = 0.018703243, train_loss = 70.03967 (0.686 sec)\n",
"INFO:tensorflow:loss = 70.03967, step = 42 (0.686 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62558\n",
"INFO:tensorflow:train_accuracy = 0.029062087, train_loss = 65.07764 (0.615 sec)\n",
"INFO:tensorflow:loss = 65.07764, step = 43 (0.615 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91277\n",
"INFO:tensorflow:train_accuracy = 0.03612903, train_loss = 65.82846 (0.523 sec)\n",
"INFO:tensorflow:loss = 65.82846, step = 44 (0.523 sec)\n",
"INFO:tensorflow:global_step/sec: 4.34446\n",
"INFO:tensorflow:train_accuracy = 0.055, train_loss = 56.317627 (0.230 sec)\n",
"INFO:tensorflow:loss = 56.317627, step = 45 (0.230 sec)\n",
"INFO:tensorflow:global_step/sec: 1.68901\n",
"INFO:tensorflow:train_accuracy = 0.032258064, train_loss = 62.233498 (0.592 sec)\n",
"INFO:tensorflow:loss = 62.233498, step = 46 (0.593 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63454\n",
"INFO:tensorflow:train_accuracy = 0.041939713, train_loss = 65.322586 (0.612 sec)\n",
"INFO:tensorflow:loss = 65.322586, step = 47 (0.611 sec)\n",
"INFO:tensorflow:global_step/sec: 1.42025\n",
"INFO:tensorflow:train_accuracy = 0.032926828, train_loss = 67.62929 (0.704 sec)\n",
"INFO:tensorflow:loss = 67.62929, step = 48 (0.704 sec)\n",
"INFO:tensorflow:global_step/sec: 1.59962\n",
"INFO:tensorflow:train_accuracy = 0.034653466, train_loss = 66.89864 (0.625 sec)\n",
"INFO:tensorflow:loss = 66.89864, step = 49 (0.624 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82216\n",
"INFO:tensorflow:train_accuracy = 0.035433073, train_loss = 60.75713 (0.549 sec)\n",
"INFO:tensorflow:loss = 60.75713, step = 50 (0.549 sec)\n",
"INFO:tensorflow:global_step/sec: 1.6328\n",
"INFO:tensorflow:train_accuracy = 0.04551201, train_loss = 61.6336 (0.613 sec)\n",
"INFO:tensorflow:loss = 61.6336, step = 51 (0.613 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69055\n",
"INFO:tensorflow:train_accuracy = 0.04172099, train_loss = 61.794777 (0.592 sec)\n",
"INFO:tensorflow:loss = 61.794777, step = 52 (0.592 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66638\n",
"INFO:tensorflow:train_accuracy = 0.036096256, train_loss = 57.94404 (0.600 sec)\n",
"INFO:tensorflow:loss = 57.94404, step = 53 (0.600 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89491\n",
"INFO:tensorflow:train_accuracy = 0.0374677, train_loss = 59.235607 (0.528 sec)\n",
"INFO:tensorflow:loss = 59.235607, step = 54 (0.528 sec)\n",
"INFO:tensorflow:global_step/sec: 1.45216\n",
"INFO:tensorflow:train_accuracy = 0.022280471, train_loss = 61.501915 (0.689 sec)\n",
"INFO:tensorflow:loss = 61.501915, step = 55 (0.689 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63537\n",
"INFO:tensorflow:train_accuracy = 0.042801555, train_loss = 57.738037 (0.611 sec)\n",
"INFO:tensorflow:loss = 57.738037, step = 56 (0.611 sec)\n",
"INFO:tensorflow:global_step/sec: 1.92095\n",
"INFO:tensorflow:train_accuracy = 0.02875817, train_loss = 57.213104 (0.521 sec)\n",
"INFO:tensorflow:loss = 57.213104, step = 57 (0.521 sec)\n",
"INFO:tensorflow:global_step/sec: 1.6438\n",
"INFO:tensorflow:train_accuracy = 0.051587302, train_loss = 54.536713 (0.608 sec)\n",
"INFO:tensorflow:loss = 54.536713, step = 58 (0.608 sec)\n",
"INFO:tensorflow:global_step/sec: 1.41297\n",
"INFO:tensorflow:train_accuracy = 0.049868766, train_loss = 55.203407 (0.708 sec)\n",
"INFO:tensorflow:loss = 55.203407, step = 59 (0.707 sec)\n",
"INFO:tensorflow:global_step/sec: 2.01815\n",
"INFO:tensorflow:train_accuracy = 0.051630434, train_loss = 54.689804 (0.496 sec)\n",
"INFO:tensorflow:loss = 54.689804, step = 60 (0.496 sec)\n",
"INFO:tensorflow:global_step/sec: 1.822\n",
"INFO:tensorflow:train_accuracy = 0.05324675, train_loss = 56.476532 (0.549 sec)\n",
"INFO:tensorflow:loss = 56.476532, step = 61 (0.549 sec)\n",
"INFO:tensorflow:global_step/sec: 1.93062\n",
"INFO:tensorflow:train_accuracy = 0.037958115, train_loss = 59.659557 (0.518 sec)\n",
"INFO:tensorflow:loss = 59.659557, step = 62 (0.518 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87153\n",
"INFO:tensorflow:train_accuracy = 0.04654088, train_loss = 59.812447 (0.534 sec)\n",
"INFO:tensorflow:loss = 59.812447, step = 63 (0.534 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89226\n",
"INFO:tensorflow:train_accuracy = 0.032258064, train_loss = 51.74657 (0.529 sec)\n",
"INFO:tensorflow:loss = 51.74657, step = 64 (0.529 sec)\n",
"INFO:tensorflow:global_step/sec: 1.81604\n",
"INFO:tensorflow:train_accuracy = 0.041612484, train_loss = 59.7442 (0.551 sec)\n",
"INFO:tensorflow:loss = 59.7442, step = 65 (0.551 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62705\n",
"INFO:tensorflow:train_accuracy = 0.020151133, train_loss = 61.53784 (0.614 sec)\n",
"INFO:tensorflow:loss = 61.53784, step = 66 (0.614 sec)\n",
"INFO:tensorflow:global_step/sec: 1.6381\n",
"INFO:tensorflow:train_accuracy = 0.040843215, train_loss = 57.524162 (0.610 sec)\n",
"INFO:tensorflow:loss = 57.524162, step = 67 (0.611 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77066\n",
"INFO:tensorflow:train_accuracy = 0.045751635, train_loss = 55.68164 (0.566 sec)\n",
"INFO:tensorflow:loss = 55.68164, step = 68 (0.565 sec)\n",
"INFO:tensorflow:global_step/sec: 1.79374\n",
"INFO:tensorflow:train_accuracy = 0.055330634, train_loss = 49.815926 (0.557 sec)\n",
"INFO:tensorflow:loss = 49.815926, step = 69 (0.557 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87635\n",
"INFO:tensorflow:train_accuracy = 0.07151819, train_loss = 57.123684 (0.532 sec)\n",
"INFO:tensorflow:loss = 57.123684, step = 70 (0.532 sec)\n",
"INFO:tensorflow:global_step/sec: 1.83293\n",
"INFO:tensorflow:train_accuracy = 0.04761905, train_loss = 55.84402 (0.546 sec)\n",
"INFO:tensorflow:loss = 55.84402, step = 71 (0.546 sec)\n",
"INFO:tensorflow:global_step/sec: 1.45229\n",
"INFO:tensorflow:train_accuracy = 0.045691907, train_loss = 57.590637 (0.688 sec)\n",
"INFO:tensorflow:loss = 57.590637, step = 72 (0.688 sec)\n",
"INFO:tensorflow:global_step/sec: 1.38967\n",
"INFO:tensorflow:train_accuracy = 0.04805195, train_loss = 57.315422 (0.720 sec)\n",
"INFO:tensorflow:loss = 57.315422, step = 73 (0.720 sec)\n",
"INFO:tensorflow:global_step/sec: 1.778\n",
"INFO:tensorflow:train_accuracy = 0.04665826, train_loss = 56.277447 (0.562 sec)\n",
"INFO:tensorflow:loss = 56.277447, step = 74 (0.562 sec)\n",
"INFO:tensorflow:global_step/sec: 1.55487\n",
"INFO:tensorflow:train_accuracy = 0.047919292, train_loss = 63.59537 (0.643 sec)\n",
"INFO:tensorflow:loss = 63.59537, step = 75 (0.644 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91505\n",
"INFO:tensorflow:train_accuracy = 0.041198503, train_loss = 58.507576 (0.522 sec)\n",
"INFO:tensorflow:loss = 58.507576, step = 76 (0.523 sec)\n",
"INFO:tensorflow:global_step/sec: 1.51039\n",
"INFO:tensorflow:train_accuracy = 0.043092523, train_loss = 54.04271 (0.662 sec)\n",
"INFO:tensorflow:loss = 54.04271, step = 77 (0.661 sec)\n",
"INFO:tensorflow:global_step/sec: 1.83353\n",
"INFO:tensorflow:train_accuracy = 0.044854883, train_loss = 55.491005 (0.546 sec)\n",
"INFO:tensorflow:loss = 55.491005, step = 78 (0.546 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74554\n",
"INFO:tensorflow:train_accuracy = 0.04045512, train_loss = 53.026646 (0.573 sec)\n",
"INFO:tensorflow:loss = 53.026646, step = 79 (0.572 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75313\n",
"INFO:tensorflow:train_accuracy = 0.03606965, train_loss = 58.148933 (0.571 sec)\n",
"INFO:tensorflow:loss = 58.148933, step = 80 (0.571 sec)\n",
"INFO:tensorflow:global_step/sec: 1.58324\n",
"INFO:tensorflow:train_accuracy = 0.043422733, train_loss = 57.046684 (0.631 sec)\n",
"INFO:tensorflow:loss = 57.046684, step = 81 (0.631 sec)\n",
"INFO:tensorflow:global_step/sec: 1.9075\n",
"INFO:tensorflow:train_accuracy = 0.08070618, train_loss = 56.546654 (0.525 sec)\n",
"INFO:tensorflow:loss = 56.546654, step = 82 (0.525 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73199\n",
"INFO:tensorflow:train_accuracy = 0.076030925, train_loss = 58.30722 (0.577 sec)\n",
"INFO:tensorflow:loss = 58.30722, step = 83 (0.577 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73954\n",
"INFO:tensorflow:train_accuracy = 0.057544757, train_loss = 53.906185 (0.575 sec)\n",
"INFO:tensorflow:loss = 53.906185, step = 84 (0.575 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69014\n",
"INFO:tensorflow:train_accuracy = 0.04767726, train_loss = 57.648247 (0.592 sec)\n",
"INFO:tensorflow:loss = 57.648247, step = 85 (0.591 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75705\n",
"INFO:tensorflow:train_accuracy = 0.060489062, train_loss = 51.53795 (0.569 sec)\n",
"INFO:tensorflow:loss = 51.53795, step = 86 (0.570 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86749\n",
"INFO:tensorflow:train_accuracy = 0.08476821, train_loss = 47.573967 (0.535 sec)\n",
"INFO:tensorflow:loss = 47.573967, step = 87 (0.535 sec)\n",
"INFO:tensorflow:global_step/sec: 1.96024\n",
"INFO:tensorflow:train_accuracy = 0.073264785, train_loss = 45.025467 (0.510 sec)\n",
"INFO:tensorflow:loss = 45.025467, step = 88 (0.510 sec)\n",
"INFO:tensorflow:global_step/sec: 1.97588\n",
"INFO:tensorflow:train_accuracy = 0.15753424, train_loss = 41.763363 (0.506 sec)\n",
"INFO:tensorflow:loss = 41.763363, step = 89 (0.506 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91981\n",
"INFO:tensorflow:train_accuracy = 0.1891892, train_loss = 47.59393 (0.521 sec)\n",
"INFO:tensorflow:loss = 47.59393, step = 90 (0.521 sec)\n",
"INFO:tensorflow:global_step/sec: 3.89724\n",
"INFO:tensorflow:train_accuracy = 0.086538464, train_loss = 44.609444 (0.256 sec)\n",
"INFO:tensorflow:loss = 44.609444, step = 91 (0.256 sec)\n",
"INFO:tensorflow:global_step/sec: 1.5191\n",
"INFO:tensorflow:train_accuracy = 0.12823834, train_loss = 44.68122 (0.658 sec)\n",
"INFO:tensorflow:loss = 44.68122, step = 92 (0.659 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66392\n",
"INFO:tensorflow:train_accuracy = 0.08635794, train_loss = 47.669533 (0.601 sec)\n",
"INFO:tensorflow:loss = 47.669533, step = 93 (0.601 sec)\n",
"INFO:tensorflow:global_step/sec: 1.56371\n",
"INFO:tensorflow:train_accuracy = 0.102998696, train_loss = 43.111664 (0.639 sec)\n",
"INFO:tensorflow:loss = 43.111664, step = 94 (0.639 sec)\n",
"INFO:tensorflow:global_step/sec: 1.42167\n",
"INFO:tensorflow:train_accuracy = 0.13284133, train_loss = 53.00408 (0.703 sec)\n",
"INFO:tensorflow:loss = 53.00408, step = 95 (0.703 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61946\n",
"INFO:tensorflow:train_accuracy = 0.18467337, train_loss = 48.228943 (0.618 sec)\n",
"INFO:tensorflow:loss = 48.228943, step = 96 (0.618 sec)\n",
"INFO:tensorflow:global_step/sec: 1.47609\n",
"INFO:tensorflow:train_accuracy = 0.14824447, train_loss = 45.144753 (0.677 sec)\n",
"INFO:tensorflow:loss = 45.144753, step = 97 (0.677 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72856\n",
"INFO:tensorflow:train_accuracy = 0.1730013, train_loss = 42.326298 (0.579 sec)\n",
"INFO:tensorflow:loss = 42.326298, step = 98 (0.579 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69093\n",
"INFO:tensorflow:train_accuracy = 0.1994852, train_loss = 42.84948 (0.591 sec)\n",
"INFO:tensorflow:loss = 42.84948, step = 99 (0.591 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63143\n",
"INFO:tensorflow:train_accuracy = 0.10394737, train_loss = 51.87351 (0.614 sec)\n",
"INFO:tensorflow:loss = 51.87351, step = 100 (0.613 sec)\n",
"INFO:tensorflow:global_step/sec: 1.46548\n",
"INFO:tensorflow:train_accuracy = 0.09625668, train_loss = 52.595783 (0.682 sec)\n",
"INFO:tensorflow:loss = 52.595783, step = 101 (0.682 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91091\n",
"INFO:tensorflow:train_accuracy = 0.13636364, train_loss = 58.338585 (0.523 sec)\n",
"INFO:tensorflow:loss = 58.338585, step = 102 (0.523 sec)\n",
"INFO:tensorflow:global_step/sec: 1.98432\n",
"INFO:tensorflow:train_accuracy = 0.17379679, train_loss = 45.606537 (0.504 sec)\n",
"INFO:tensorflow:loss = 45.606537, step = 103 (0.504 sec)\n",
"INFO:tensorflow:global_step/sec: 1.96924\n",
"INFO:tensorflow:train_accuracy = 0.19121447, train_loss = 51.812874 (0.508 sec)\n",
"INFO:tensorflow:loss = 51.812874, step = 104 (0.507 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64311\n",
"INFO:tensorflow:train_accuracy = 0.17513369, train_loss = 49.92538 (0.608 sec)\n",
"INFO:tensorflow:loss = 49.92538, step = 105 (0.608 sec)\n",
"INFO:tensorflow:global_step/sec: 1.92304\n",
"INFO:tensorflow:train_accuracy = 0.043984476, train_loss = 48.259773 (0.520 sec)\n",
"INFO:tensorflow:loss = 48.259773, step = 106 (0.521 sec)\n",
"INFO:tensorflow:global_step/sec: 1.83366\n",
"INFO:tensorflow:train_accuracy = 0.12840466, train_loss = 49.470116 (0.545 sec)\n",
"INFO:tensorflow:loss = 49.470116, step = 107 (0.545 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91397\n",
"INFO:tensorflow:train_accuracy = 0.16821192, train_loss = 44.4474 (0.523 sec)\n",
"INFO:tensorflow:loss = 44.4474, step = 108 (0.523 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88424\n",
"INFO:tensorflow:train_accuracy = 0.12549537, train_loss = 41.46634 (0.531 sec)\n",
"INFO:tensorflow:loss = 41.46634, step = 109 (0.531 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82204\n",
"INFO:tensorflow:train_accuracy = 0.14138818, train_loss = 49.29923 (0.549 sec)\n",
"INFO:tensorflow:loss = 49.29923, step = 110 (0.549 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8114\n",
"INFO:tensorflow:train_accuracy = 0.12827225, train_loss = 45.673637 (0.552 sec)\n",
"INFO:tensorflow:loss = 45.673637, step = 111 (0.552 sec)\n",
"INFO:tensorflow:global_step/sec: 1.83569\n",
"INFO:tensorflow:train_accuracy = 0.10981912, train_loss = 44.434998 (0.545 sec)\n",
"INFO:tensorflow:loss = 44.434998, step = 112 (0.545 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63053\n",
"INFO:tensorflow:train_accuracy = 0.06786171, train_loss = 44.19197 (0.613 sec)\n",
"INFO:tensorflow:loss = 44.19197, step = 113 (0.613 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7586\n",
"INFO:tensorflow:train_accuracy = 0.08025478, train_loss = 55.78295 (0.568 sec)\n",
"INFO:tensorflow:loss = 55.78295, step = 114 (0.568 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88541\n",
"INFO:tensorflow:train_accuracy = 0.1618799, train_loss = 41.771008 (0.531 sec)\n",
"INFO:tensorflow:loss = 41.771008, step = 115 (0.531 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66683\n",
"INFO:tensorflow:train_accuracy = 0.10909091, train_loss = 41.753586 (0.600 sec)\n",
"INFO:tensorflow:loss = 41.753586, step = 116 (0.599 sec)\n",
"INFO:tensorflow:global_step/sec: 1.92286\n",
"INFO:tensorflow:train_accuracy = 0.118983954, train_loss = 39.700058 (0.520 sec)\n",
"INFO:tensorflow:loss = 39.700058, step = 117 (0.521 sec)\n",
"INFO:tensorflow:global_step/sec: 1.68828\n",
"INFO:tensorflow:train_accuracy = 0.1334214, train_loss = 39.467422 (0.593 sec)\n",
"INFO:tensorflow:loss = 39.467422, step = 118 (0.593 sec)\n",
"INFO:tensorflow:global_step/sec: 1.50782\n",
"INFO:tensorflow:train_accuracy = 0.1425031, train_loss = 56.011097 (0.663 sec)\n",
"INFO:tensorflow:loss = 56.011097, step = 119 (0.663 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63566\n",
"INFO:tensorflow:train_accuracy = 0.19502617, train_loss = 43.18344 (0.611 sec)\n",
"INFO:tensorflow:loss = 43.18344, step = 120 (0.612 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85435\n",
"INFO:tensorflow:train_accuracy = 0.24402516, train_loss = 47.57972 (0.540 sec)\n",
"INFO:tensorflow:loss = 47.57972, step = 121 (0.540 sec)\n",
"INFO:tensorflow:global_step/sec: 1.49704\n",
"INFO:tensorflow:train_accuracy = 0.16814159, train_loss = 51.164787 (0.668 sec)\n",
"INFO:tensorflow:loss = 51.164787, step = 122 (0.667 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77849\n",
"INFO:tensorflow:train_accuracy = 0.2496749, train_loss = 39.28454 (0.562 sec)\n",
"INFO:tensorflow:loss = 39.28454, step = 123 (0.563 sec)\n",
"INFO:tensorflow:global_step/sec: 1.49858\n",
"INFO:tensorflow:train_accuracy = 0.14791404, train_loss = 44.27699 (0.667 sec)\n",
"INFO:tensorflow:loss = 44.27699, step = 124 (0.667 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62974\n",
"INFO:tensorflow:train_accuracy = 0.19375, train_loss = 42.398117 (0.614 sec)\n",
"INFO:tensorflow:loss = 42.398117, step = 125 (0.614 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73934\n",
"INFO:tensorflow:train_accuracy = 0.23390275, train_loss = 39.59218 (0.575 sec)\n",
"INFO:tensorflow:loss = 39.59218, step = 126 (0.575 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88901\n",
"INFO:tensorflow:train_accuracy = 0.24320827, train_loss = 37.307465 (0.529 sec)\n",
"INFO:tensorflow:loss = 37.307465, step = 127 (0.529 sec)\n",
"INFO:tensorflow:global_step/sec: 1.57988\n",
"INFO:tensorflow:train_accuracy = 0.22250316, train_loss = 40.344303 (0.633 sec)\n",
"INFO:tensorflow:loss = 40.344303, step = 128 (0.633 sec)\n",
"INFO:tensorflow:global_step/sec: 1.58625\n",
"INFO:tensorflow:train_accuracy = 0.21744472, train_loss = 42.249992 (0.630 sec)\n",
"INFO:tensorflow:loss = 42.249992, step = 129 (0.630 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74825\n",
"INFO:tensorflow:train_accuracy = 0.26206, train_loss = 40.40354 (0.572 sec)\n",
"INFO:tensorflow:loss = 40.40354, step = 130 (0.573 sec)\n",
"INFO:tensorflow:global_step/sec: 2.00801\n",
"INFO:tensorflow:train_accuracy = 0.24590164, train_loss = 39.785133 (0.498 sec)\n",
"INFO:tensorflow:loss = 39.785133, step = 131 (0.498 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72104\n",
"INFO:tensorflow:train_accuracy = 0.23341836, train_loss = 40.44952 (0.581 sec)\n",
"INFO:tensorflow:loss = 40.44952, step = 132 (0.581 sec)\n",
"INFO:tensorflow:global_step/sec: 1.93401\n",
"INFO:tensorflow:train_accuracy = 0.23521683, train_loss = 36.154213 (0.516 sec)\n",
"INFO:tensorflow:loss = 36.154213, step = 133 (0.516 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75269\n",
"INFO:tensorflow:train_accuracy = 0.24784216, train_loss = 40.154743 (0.571 sec)\n",
"INFO:tensorflow:loss = 40.154743, step = 134 (0.572 sec)\n",
"INFO:tensorflow:global_step/sec: 1.94338\n",
"INFO:tensorflow:train_accuracy = 0.23834887, train_loss = 34.376724 (0.514 sec)\n",
"INFO:tensorflow:loss = 34.376724, step = 135 (0.514 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85933\n",
"INFO:tensorflow:train_accuracy = 0.2868526, train_loss = 32.17231 (0.538 sec)\n",
"INFO:tensorflow:loss = 32.17231, step = 136 (0.538 sec)\n",
"INFO:tensorflow:global_step/sec: 4.4516\n",
"INFO:tensorflow:train_accuracy = 0.15656565, train_loss = 52.824837 (0.224 sec)\n",
"INFO:tensorflow:loss = 52.824837, step = 137 (0.224 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66861\n",
"INFO:tensorflow:train_accuracy = 0.26444158, train_loss = 35.6593 (0.599 sec)\n",
"INFO:tensorflow:loss = 35.6593, step = 138 (0.599 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65073\n",
"INFO:tensorflow:train_accuracy = 0.21509434, train_loss = 39.754944 (0.606 sec)\n",
"INFO:tensorflow:loss = 39.754944, step = 139 (0.606 sec)\n",
"INFO:tensorflow:global_step/sec: 1.4316\n",
"INFO:tensorflow:train_accuracy = 0.22331288, train_loss = 47.32818 (0.699 sec)\n",
"INFO:tensorflow:loss = 47.32818, step = 140 (0.699 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82721\n",
"INFO:tensorflow:train_accuracy = 0.21693122, train_loss = 34.57155 (0.547 sec)\n",
"INFO:tensorflow:loss = 34.57155, step = 141 (0.547 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85692\n",
"INFO:tensorflow:train_accuracy = 0.20330368, train_loss = 39.584354 (0.538 sec)\n",
"INFO:tensorflow:loss = 39.584354, step = 142 (0.538 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63098\n",
"INFO:tensorflow:train_accuracy = 0.2535211, train_loss = 35.705524 (0.614 sec)\n",
"INFO:tensorflow:loss = 35.705524, step = 143 (0.614 sec)\n",
"INFO:tensorflow:global_step/sec: 1.58955\n",
"INFO:tensorflow:train_accuracy = 0.28627452, train_loss = 32.651062 (0.629 sec)\n",
"INFO:tensorflow:loss = 32.651062, step = 144 (0.629 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65532\n",
"INFO:tensorflow:train_accuracy = 0.30839896, train_loss = 31.46363 (0.604 sec)\n",
"INFO:tensorflow:loss = 31.46363, step = 145 (0.604 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62624\n",
"INFO:tensorflow:train_accuracy = 0.27673796, train_loss = 30.687359 (0.615 sec)\n",
"INFO:tensorflow:loss = 30.687359, step = 146 (0.615 sec)\n",
"INFO:tensorflow:global_step/sec: 1.45792\n",
"INFO:tensorflow:train_accuracy = 0.2916129, train_loss = 35.135002 (0.686 sec)\n",
"INFO:tensorflow:loss = 35.135002, step = 147 (0.686 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71751\n",
"INFO:tensorflow:train_accuracy = 0.3188406, train_loss = 31.725975 (0.582 sec)\n",
"INFO:tensorflow:loss = 31.725975, step = 148 (0.582 sec)\n",
"INFO:tensorflow:global_step/sec: 1.724\n",
"INFO:tensorflow:train_accuracy = 0.31233245, train_loss = 31.470583 (0.580 sec)\n",
"INFO:tensorflow:loss = 31.470583, step = 149 (0.580 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63232\n",
"INFO:tensorflow:train_accuracy = 0.2364096, train_loss = 38.853527 (0.612 sec)\n",
"INFO:tensorflow:loss = 38.853527, step = 150 (0.612 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88349\n",
"INFO:tensorflow:train_accuracy = 0.30789474, train_loss = 32.435463 (0.531 sec)\n",
"INFO:tensorflow:loss = 32.435463, step = 151 (0.532 sec)\n",
"INFO:tensorflow:global_step/sec: 1.92772\n",
"INFO:tensorflow:train_accuracy = 0.31903484, train_loss = 33.201637 (0.518 sec)\n",
"INFO:tensorflow:loss = 33.201637, step = 152 (0.518 sec)\n",
"INFO:tensorflow:global_step/sec: 1.99014\n",
"INFO:tensorflow:train_accuracy = 0.28917196, train_loss = 33.46064 (0.503 sec)\n",
"INFO:tensorflow:loss = 33.46064, step = 153 (0.503 sec)\n",
"INFO:tensorflow:global_step/sec: 1.98867\n",
"INFO:tensorflow:train_accuracy = 0.31530344, train_loss = 31.209377 (0.503 sec)\n",
"INFO:tensorflow:loss = 31.209377, step = 154 (0.503 sec)\n",
"INFO:tensorflow:global_step/sec: 1.728\n",
"INFO:tensorflow:train_accuracy = 0.26441103, train_loss = 36.82923 (0.579 sec)\n",
"INFO:tensorflow:loss = 36.82923, step = 155 (0.579 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82308\n",
"INFO:tensorflow:train_accuracy = 0.314888, train_loss = 30.534271 (0.548 sec)\n",
"INFO:tensorflow:loss = 30.534271, step = 156 (0.548 sec)\n",
"INFO:tensorflow:global_step/sec: 1.83945\n",
"INFO:tensorflow:train_accuracy = 0.38, train_loss = 29.77198 (0.544 sec)\n",
"INFO:tensorflow:loss = 29.77198, step = 157 (0.544 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63486\n",
"INFO:tensorflow:train_accuracy = 0.33375958, train_loss = 34.230366 (0.612 sec)\n",
"INFO:tensorflow:loss = 34.230366, step = 158 (0.611 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88538\n",
"INFO:tensorflow:train_accuracy = 0.3276316, train_loss = 32.790657 (0.531 sec)\n",
"INFO:tensorflow:loss = 32.790657, step = 159 (0.531 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73614\n",
"INFO:tensorflow:train_accuracy = 0.3159948, train_loss = 30.283646 (0.575 sec)\n",
"INFO:tensorflow:loss = 30.283646, step = 160 (0.575 sec)\n",
"INFO:tensorflow:global_step/sec: 2.04173\n",
"INFO:tensorflow:train_accuracy = 0.32207793, train_loss = 29.865269 (0.490 sec)\n",
"INFO:tensorflow:loss = 29.865269, step = 161 (0.491 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87317\n",
"INFO:tensorflow:train_accuracy = 0.33643618, train_loss = 27.693954 (0.534 sec)\n",
"INFO:tensorflow:loss = 27.693954, step = 162 (0.534 sec)\n",
"INFO:tensorflow:global_step/sec: 1.95108\n",
"INFO:tensorflow:train_accuracy = 0.31893265, train_loss = 34.291832 (0.513 sec)\n",
"INFO:tensorflow:loss = 34.291832, step = 163 (0.513 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80029\n",
"INFO:tensorflow:train_accuracy = 0.30331126, train_loss = 29.584002 (0.555 sec)\n",
"INFO:tensorflow:loss = 29.584002, step = 164 (0.555 sec)\n",
"INFO:tensorflow:global_step/sec: 1.52845\n",
"INFO:tensorflow:train_accuracy = 0.31698114, train_loss = 33.162903 (0.654 sec)\n",
"INFO:tensorflow:loss = 33.162903, step = 165 (0.654 sec)\n",
"INFO:tensorflow:global_step/sec: 1.93201\n",
"INFO:tensorflow:train_accuracy = 0.35923567, train_loss = 30.956665 (0.518 sec)\n",
"INFO:tensorflow:loss = 30.956665, step = 166 (0.518 sec)\n",
"INFO:tensorflow:global_step/sec: 2.12795\n",
"INFO:tensorflow:train_accuracy = 0.25, train_loss = 35.47502 (0.470 sec)\n",
"INFO:tensorflow:loss = 35.47502, step = 167 (0.470 sec)\n",
"INFO:tensorflow:global_step/sec: 1.52354\n",
"INFO:tensorflow:train_accuracy = 0.34987894, train_loss = 42.086166 (0.656 sec)\n",
"INFO:tensorflow:loss = 42.086166, step = 168 (0.657 sec)\n",
"INFO:tensorflow:global_step/sec: 1.57279\n",
"INFO:tensorflow:train_accuracy = 0.3483444, train_loss = 28.382399 (0.636 sec)\n",
"INFO:tensorflow:loss = 28.382399, step = 169 (0.636 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63977\n",
"INFO:tensorflow:train_accuracy = 0.36713734, train_loss = 37.42907 (0.610 sec)\n",
"INFO:tensorflow:loss = 37.42907, step = 170 (0.610 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74009\n",
"INFO:tensorflow:train_accuracy = 0.26986128, train_loss = 32.643154 (0.575 sec)\n",
"INFO:tensorflow:loss = 32.643154, step = 171 (0.576 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80354\n",
"INFO:tensorflow:train_accuracy = 0.28188777, train_loss = 31.717594 (0.554 sec)\n",
"INFO:tensorflow:loss = 31.717594, step = 172 (0.554 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61291\n",
"INFO:tensorflow:train_accuracy = 0.26641884, train_loss = 37.64123 (0.620 sec)\n",
"INFO:tensorflow:loss = 37.64123, step = 173 (0.620 sec)\n",
"INFO:tensorflow:global_step/sec: 1.546\n",
"INFO:tensorflow:train_accuracy = 0.317662, train_loss = 35.836884 (0.647 sec)\n",
"INFO:tensorflow:loss = 35.836884, step = 174 (0.646 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73688\n",
"INFO:tensorflow:train_accuracy = 0.30264816, train_loss = 32.92005 (0.576 sec)\n",
"INFO:tensorflow:loss = 32.92005, step = 175 (0.576 sec)\n",
"INFO:tensorflow:global_step/sec: 1.92199\n",
"INFO:tensorflow:train_accuracy = 0.34986946, train_loss = 29.878262 (0.520 sec)\n",
"INFO:tensorflow:loss = 29.878262, step = 176 (0.521 sec)\n",
"INFO:tensorflow:global_step/sec: 1.70435\n",
"INFO:tensorflow:train_accuracy = 0.34362933, train_loss = 33.081245 (0.586 sec)\n",
"INFO:tensorflow:loss = 33.081245, step = 177 (0.586 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87234\n",
"INFO:tensorflow:train_accuracy = 0.2918848, train_loss = 30.92514 (0.534 sec)\n",
"INFO:tensorflow:loss = 30.92514, step = 178 (0.534 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86119\n",
"INFO:tensorflow:train_accuracy = 0.2534854, train_loss = 37.991367 (0.538 sec)\n",
"INFO:tensorflow:loss = 37.991367, step = 179 (0.538 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80532\n",
"INFO:tensorflow:train_accuracy = 0.35353535, train_loss = 35.757095 (0.554 sec)\n",
"INFO:tensorflow:loss = 35.757095, step = 180 (0.554 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80514\n",
"INFO:tensorflow:train_accuracy = 0.27919462, train_loss = 39.81878 (0.554 sec)\n",
"INFO:tensorflow:loss = 39.81878, step = 181 (0.554 sec)\n",
"INFO:tensorflow:global_step/sec: 1.92906\n",
"INFO:tensorflow:train_accuracy = 0.35914552, train_loss = 26.343407 (0.518 sec)\n",
"INFO:tensorflow:loss = 26.343407, step = 182 (0.518 sec)\n",
"INFO:tensorflow:global_step/sec: 4.02599\n",
"INFO:tensorflow:train_accuracy = 0.29577464, train_loss = 39.315094 (0.248 sec)\n",
"INFO:tensorflow:loss = 39.315094, step = 183 (0.248 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62647\n",
"INFO:tensorflow:train_accuracy = 0.31675392, train_loss = 30.443237 (0.615 sec)\n",
"INFO:tensorflow:loss = 30.443237, step = 184 (0.615 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87773\n",
"INFO:tensorflow:train_accuracy = 0.29861984, train_loss = 34.694077 (0.533 sec)\n",
"INFO:tensorflow:loss = 34.694077, step = 185 (0.533 sec)\n",
"INFO:tensorflow:global_step/sec: 1.43724\n",
"INFO:tensorflow:train_accuracy = 0.25653797, train_loss = 44.484352 (0.696 sec)\n",
"INFO:tensorflow:loss = 44.484352, step = 186 (0.695 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63972\n",
"INFO:tensorflow:train_accuracy = 0.23, train_loss = 47.01816 (0.610 sec)\n",
"INFO:tensorflow:loss = 47.01816, step = 187 (0.610 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61856\n",
"INFO:tensorflow:train_accuracy = 0.29808918, train_loss = 34.06806 (0.618 sec)\n",
"INFO:tensorflow:loss = 34.06806, step = 188 (0.618 sec)\n",
"INFO:tensorflow:global_step/sec: 1.78551\n",
"INFO:tensorflow:train_accuracy = 0.31428573, train_loss = 32.31922 (0.560 sec)\n",
"INFO:tensorflow:loss = 32.31922, step = 189 (0.560 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69741\n",
"INFO:tensorflow:train_accuracy = 0.27906978, train_loss = 32.944515 (0.589 sec)\n",
"INFO:tensorflow:loss = 32.944515, step = 190 (0.589 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73534\n",
"INFO:tensorflow:train_accuracy = 0.31697613, train_loss = 33.505646 (0.576 sec)\n",
"INFO:tensorflow:loss = 33.505646, step = 191 (0.576 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63272\n",
"INFO:tensorflow:train_accuracy = 0.35402906, train_loss = 29.834244 (0.613 sec)\n",
"INFO:tensorflow:loss = 29.834244, step = 192 (0.613 sec)\n",
"INFO:tensorflow:global_step/sec: 1.44219\n",
"INFO:tensorflow:train_accuracy = 0.2745358, train_loss = 33.38021 (0.693 sec)\n",
"INFO:tensorflow:loss = 33.38021, step = 193 (0.693 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65161\n",
"INFO:tensorflow:train_accuracy = 0.24146982, train_loss = 46.30552 (0.605 sec)\n",
"INFO:tensorflow:loss = 46.30552, step = 194 (0.605 sec)\n",
"INFO:tensorflow:global_step/sec: 1.94391\n",
"INFO:tensorflow:train_accuracy = 0.28515112, train_loss = 46.706978 (0.514 sec)\n",
"INFO:tensorflow:loss = 46.706978, step = 195 (0.514 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90761\n",
"INFO:tensorflow:train_accuracy = 0.28480205, train_loss = 44.374542 (0.525 sec)\n",
"INFO:tensorflow:loss = 44.374542, step = 196 (0.525 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61826\n",
"INFO:tensorflow:train_accuracy = 0.29903978, train_loss = 38.49812 (0.618 sec)\n",
"INFO:tensorflow:loss = 38.49812, step = 197 (0.618 sec)\n",
"INFO:tensorflow:global_step/sec: 2.12319\n",
"INFO:tensorflow:train_accuracy = 0.26877472, train_loss = 43.2146 (0.471 sec)\n",
"INFO:tensorflow:loss = 43.2146, step = 198 (0.471 sec)\n",
"INFO:tensorflow:Saving checkpoints for 200 into asr-quartznet/model.ckpt.\n",
"INFO:tensorflow:Calling model_fn.\n",
"WARNING:tensorflow:From /home/husein/malaya-speech/malaya_speech/train/model/ctc/metrics.py:58: The name tf.metrics.mean is deprecated. Please use tf.compat.v1.metrics.mean instead.\n",
"\n",
"WARNING:tensorflow:From /home/husein/malaya-speech/malaya_speech/train/model/ctc/metrics.py:106: The name tf.regex_replace is deprecated. Please use tf.strings.regex_replace instead.\n",
"\n",
"INFO:tensorflow:Done calling model_fn.\n",
"INFO:tensorflow:Starting evaluation at 2020-10-25T18:56:33Z\n",
"INFO:tensorflow:Graph was finalized.\n",
"INFO:tensorflow:Restoring parameters from asr-quartznet/model.ckpt-200\n",
"INFO:tensorflow:Running local_init_op.\n",
"INFO:tensorflow:Done running local_init_op.\n",
"INFO:tensorflow:Evaluation [1/10]\n",
"INFO:tensorflow:Evaluation [2/10]\n",
"INFO:tensorflow:Evaluation [3/10]\n",
"INFO:tensorflow:Evaluation [4/10]\n",
"INFO:tensorflow:Evaluation [5/10]\n",
"INFO:tensorflow:Evaluation [6/10]\n",
"INFO:tensorflow:Evaluation [7/10]\n",
"INFO:tensorflow:Evaluation [8/10]\n",
"INFO:tensorflow:Evaluation [9/10]\n",
"INFO:tensorflow:Evaluation [10/10]\n",
"INFO:tensorflow:Finished evaluation at 2020-10-25-18:56:42\n",
"INFO:tensorflow:Saving dict for global step 200: WER = 1.0, accuracy = 0.21779141, global_step = 200, loss = 40.178703\n",
"INFO:tensorflow:Saving 'checkpoint_path' summary for global step 200: asr-quartznet/model.ckpt-200\n",
"INFO:tensorflow:global_step/sec: 0.0526832\n",
"INFO:tensorflow:train_accuracy = 0.26996198, train_loss = 41.80652 (18.981 sec)\n",
"INFO:tensorflow:loss = 41.80652, step = 199 (18.980 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8924\n",
"INFO:tensorflow:train_accuracy = 0.22721438, train_loss = 42.77658 (0.530 sec)\n",
"INFO:tensorflow:loss = 42.77658, step = 200 (0.530 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8917\n",
"INFO:tensorflow:train_accuracy = 0.19897304, train_loss = 40.37663 (0.528 sec)\n",
"INFO:tensorflow:loss = 40.37663, step = 201 (0.527 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85587\n",
"INFO:tensorflow:train_accuracy = 0.24601063, train_loss = 41.40363 (0.539 sec)\n",
"INFO:tensorflow:loss = 41.40363, step = 202 (0.540 sec)\n",
"INFO:tensorflow:global_step/sec: 1.76385\n",
"INFO:tensorflow:train_accuracy = 0.18607764, train_loss = 42.333603 (0.567 sec)\n",
"INFO:tensorflow:loss = 42.333603, step = 203 (0.567 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62013\n",
"INFO:tensorflow:train_accuracy = 0.23433584, train_loss = 48.916977 (0.617 sec)\n",
"INFO:tensorflow:loss = 48.916977, step = 204 (0.617 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82314\n",
"INFO:tensorflow:train_accuracy = 0.19458763, train_loss = 43.97137 (0.549 sec)\n",
"INFO:tensorflow:loss = 43.97137, step = 205 (0.549 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77671\n",
"INFO:tensorflow:train_accuracy = 0.27931488, train_loss = 37.46124 (0.563 sec)\n",
"INFO:tensorflow:loss = 37.46124, step = 206 (0.562 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88019\n",
"INFO:tensorflow:train_accuracy = 0.2782956, train_loss = 35.177227 (0.532 sec)\n",
"INFO:tensorflow:loss = 35.177227, step = 207 (0.532 sec)\n",
"INFO:tensorflow:global_step/sec: 2.02307\n",
"INFO:tensorflow:train_accuracy = 0.28460544, train_loss = 39.268867 (0.494 sec)\n",
"INFO:tensorflow:loss = 39.268867, step = 208 (0.494 sec)\n",
"INFO:tensorflow:global_step/sec: 1.943\n",
"INFO:tensorflow:train_accuracy = 0.21279374, train_loss = 42.26107 (0.514 sec)\n",
"INFO:tensorflow:loss = 42.26107, step = 209 (0.514 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7171\n",
"INFO:tensorflow:train_accuracy = 0.3138021, train_loss = 36.207676 (0.583 sec)\n",
"INFO:tensorflow:loss = 36.207676, step = 210 (0.583 sec)\n",
"INFO:tensorflow:global_step/sec: 1.54134\n",
"INFO:tensorflow:train_accuracy = 0.30994898, train_loss = 49.0229 (0.649 sec)\n",
"INFO:tensorflow:loss = 49.0229, step = 211 (0.649 sec)\n",
"INFO:tensorflow:global_step/sec: 2.00029\n",
"INFO:tensorflow:train_accuracy = 0.3197452, train_loss = 39.306816 (0.500 sec)\n",
"INFO:tensorflow:loss = 39.306816, step = 212 (0.501 sec)\n",
"INFO:tensorflow:global_step/sec: 2.0439\n",
"INFO:tensorflow:train_accuracy = 0.23803526, train_loss = 43.07305 (0.489 sec)\n",
"INFO:tensorflow:loss = 43.07305, step = 213 (0.489 sec)\n",
"INFO:tensorflow:global_step/sec: 1.95673\n",
"INFO:tensorflow:train_accuracy = 0.26501307, train_loss = 44.222145 (0.511 sec)\n",
"INFO:tensorflow:loss = 44.222145, step = 214 (0.511 sec)\n",
"INFO:tensorflow:global_step/sec: 1.51334\n",
"INFO:tensorflow:train_accuracy = 0.26136363, train_loss = 49.082176 (0.661 sec)\n",
"INFO:tensorflow:loss = 49.082176, step = 215 (0.661 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65012\n",
"INFO:tensorflow:train_accuracy = 0.28214732, train_loss = 42.06714 (0.606 sec)\n",
"INFO:tensorflow:loss = 42.06714, step = 216 (0.606 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74942\n",
"INFO:tensorflow:train_accuracy = 0.31882203, train_loss = 38.444412 (0.571 sec)\n",
"INFO:tensorflow:loss = 38.444412, step = 217 (0.571 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86991\n",
"INFO:tensorflow:train_accuracy = 0.022164276, train_loss = 39.24765 (0.535 sec)\n",
"INFO:tensorflow:loss = 39.24765, step = 218 (0.536 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73249\n",
"INFO:tensorflow:train_accuracy = 0.03018868, train_loss = 45.537304 (0.577 sec)\n",
"INFO:tensorflow:loss = 45.537304, step = 219 (0.577 sec)\n",
"INFO:tensorflow:global_step/sec: 1.5576\n",
"INFO:tensorflow:train_accuracy = 0.060225848, train_loss = 51.820942 (0.642 sec)\n",
"INFO:tensorflow:loss = 51.820942, step = 220 (0.642 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85552\n",
"INFO:tensorflow:train_accuracy = 0.20099255, train_loss = 44.096226 (0.539 sec)\n",
"INFO:tensorflow:loss = 44.096226, step = 221 (0.539 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85614\n",
"INFO:tensorflow:train_accuracy = 0.12273902, train_loss = 39.22112 (0.538 sec)\n",
"INFO:tensorflow:loss = 39.22112, step = 222 (0.538 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71961\n",
"INFO:tensorflow:train_accuracy = 0.26383525, train_loss = 39.19892 (0.582 sec)\n",
"INFO:tensorflow:loss = 39.19892, step = 223 (0.582 sec)\n",
"INFO:tensorflow:global_step/sec: 1.96429\n",
"INFO:tensorflow:train_accuracy = 0.29435483, train_loss = 36.749752 (0.509 sec)\n",
"INFO:tensorflow:loss = 36.749752, step = 224 (0.509 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73755\n",
"INFO:tensorflow:train_accuracy = 0.258512, train_loss = 41.567444 (0.575 sec)\n",
"INFO:tensorflow:loss = 41.567444, step = 225 (0.575 sec)\n",
"INFO:tensorflow:global_step/sec: 1.94663\n",
"INFO:tensorflow:train_accuracy = 0.28981724, train_loss = 35.720646 (0.513 sec)\n",
"INFO:tensorflow:loss = 35.720646, step = 226 (0.513 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87749\n",
"INFO:tensorflow:train_accuracy = 0.31989598, train_loss = 40.33634 (0.533 sec)\n",
"INFO:tensorflow:loss = 40.33634, step = 227 (0.534 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87814\n",
"INFO:tensorflow:train_accuracy = 0.2887417, train_loss = 34.973457 (0.532 sec)\n",
"INFO:tensorflow:loss = 34.973457, step = 228 (0.532 sec)\n",
"INFO:tensorflow:global_step/sec: 4.01903\n",
"INFO:tensorflow:train_accuracy = 0.2027027, train_loss = 45.372826 (0.249 sec)\n",
"INFO:tensorflow:loss = 45.372826, step = 229 (0.249 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72971\n",
"INFO:tensorflow:train_accuracy = 0.27882037, train_loss = 33.800964 (0.578 sec)\n",
"INFO:tensorflow:loss = 33.800964, step = 230 (0.579 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77013\n",
"INFO:tensorflow:train_accuracy = 0.3465982, train_loss = 34.558075 (0.565 sec)\n",
"INFO:tensorflow:loss = 34.558075, step = 231 (0.565 sec)\n",
"INFO:tensorflow:global_step/sec: 1.40963\n",
"INFO:tensorflow:train_accuracy = 0.2514758, train_loss = 52.41411 (0.709 sec)\n",
"INFO:tensorflow:loss = 52.41411, step = 232 (0.709 sec)\n",
"INFO:tensorflow:global_step/sec: 1.59594\n",
"INFO:tensorflow:train_accuracy = 0.31113955, train_loss = 34.937798 (0.627 sec)\n",
"INFO:tensorflow:loss = 34.937798, step = 233 (0.627 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65103\n",
"INFO:tensorflow:train_accuracy = 0.22179487, train_loss = 34.390617 (0.606 sec)\n",
"INFO:tensorflow:loss = 34.390617, step = 234 (0.606 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82771\n",
"INFO:tensorflow:train_accuracy = 0.21495327, train_loss = 31.22717 (0.547 sec)\n",
"INFO:tensorflow:loss = 31.22717, step = 235 (0.547 sec)\n",
"INFO:tensorflow:global_step/sec: 1.68596\n",
"INFO:tensorflow:train_accuracy = 0.18057285, train_loss = 37.016476 (0.593 sec)\n",
"INFO:tensorflow:loss = 37.016476, step = 236 (0.593 sec)\n",
"INFO:tensorflow:global_step/sec: 1.67921\n",
"INFO:tensorflow:train_accuracy = 0.22685789, train_loss = 34.21765 (0.595 sec)\n",
"INFO:tensorflow:loss = 34.21765, step = 237 (0.595 sec)\n",
"INFO:tensorflow:global_step/sec: 1.97538\n",
"INFO:tensorflow:train_accuracy = 0.37516513, train_loss = 28.716593 (0.506 sec)\n",
"INFO:tensorflow:loss = 28.716593, step = 238 (0.506 sec)\n",
"INFO:tensorflow:global_step/sec: 1.46538\n",
"INFO:tensorflow:train_accuracy = 0.31117022, train_loss = 31.59543 (0.683 sec)\n",
"INFO:tensorflow:loss = 31.59543, step = 239 (0.683 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63624\n",
"INFO:tensorflow:train_accuracy = 0.29761904, train_loss = 29.687534 (0.611 sec)\n",
"INFO:tensorflow:loss = 29.687534, step = 240 (0.611 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8397\n",
"INFO:tensorflow:train_accuracy = 0.32594523, train_loss = 30.166416 (0.544 sec)\n",
"INFO:tensorflow:loss = 30.166416, step = 241 (0.544 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90852\n",
"INFO:tensorflow:train_accuracy = 0.3671875, train_loss = 28.721672 (0.524 sec)\n",
"INFO:tensorflow:loss = 28.721672, step = 242 (0.524 sec)\n",
"INFO:tensorflow:global_step/sec: 1.93524\n",
"INFO:tensorflow:train_accuracy = 0.27875507, train_loss = 28.656216 (0.517 sec)\n",
"INFO:tensorflow:loss = 28.656216, step = 243 (0.517 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88867\n",
"INFO:tensorflow:train_accuracy = 0.37876803, train_loss = 31.06945 (0.529 sec)\n",
"INFO:tensorflow:loss = 31.06945, step = 244 (0.529 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85691\n",
"INFO:tensorflow:train_accuracy = 0.35240573, train_loss = 29.514236 (0.539 sec)\n",
"INFO:tensorflow:loss = 29.514236, step = 245 (0.539 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85378\n",
"INFO:tensorflow:train_accuracy = 0.3746803, train_loss = 29.492529 (0.539 sec)\n",
"INFO:tensorflow:loss = 29.492529, step = 246 (0.539 sec)\n",
"INFO:tensorflow:global_step/sec: 1.84276\n",
"INFO:tensorflow:train_accuracy = 0.33748442, train_loss = 34.97162 (0.543 sec)\n",
"INFO:tensorflow:loss = 34.97162, step = 247 (0.543 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8266\n",
"INFO:tensorflow:train_accuracy = 0.40583554, train_loss = 25.206081 (0.548 sec)\n",
"INFO:tensorflow:loss = 25.206081, step = 248 (0.548 sec)\n",
"INFO:tensorflow:global_step/sec: 1.79071\n",
"INFO:tensorflow:train_accuracy = 0.34304637, train_loss = 31.013603 (0.559 sec)\n",
"INFO:tensorflow:loss = 31.013603, step = 249 (0.559 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61468\n",
"INFO:tensorflow:train_accuracy = 0.33503836, train_loss = 33.5477 (0.619 sec)\n",
"INFO:tensorflow:loss = 33.5477, step = 250 (0.619 sec)\n",
"INFO:tensorflow:global_step/sec: 1.96333\n",
"INFO:tensorflow:train_accuracy = 0.40653595, train_loss = 25.481493 (0.509 sec)\n",
"INFO:tensorflow:loss = 25.481493, step = 251 (0.509 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80131\n",
"INFO:tensorflow:train_accuracy = 0.3625498, train_loss = 25.428741 (0.555 sec)\n",
"INFO:tensorflow:loss = 25.428741, step = 252 (0.556 sec)\n",
"INFO:tensorflow:global_step/sec: 2.0117\n",
"INFO:tensorflow:train_accuracy = 0.2965879, train_loss = 28.844797 (0.497 sec)\n",
"INFO:tensorflow:loss = 28.844797, step = 253 (0.497 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87988\n",
"INFO:tensorflow:train_accuracy = 0.41145834, train_loss = 26.052343 (0.532 sec)\n",
"INFO:tensorflow:loss = 26.052343, step = 254 (0.533 sec)\n",
"INFO:tensorflow:global_step/sec: 2.03466\n",
"INFO:tensorflow:train_accuracy = 0.32980132, train_loss = 27.425585 (0.491 sec)\n",
"INFO:tensorflow:loss = 27.425585, step = 255 (0.491 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65396\n",
"INFO:tensorflow:train_accuracy = 0.35516372, train_loss = 31.487682 (0.604 sec)\n",
"INFO:tensorflow:loss = 31.487682, step = 256 (0.604 sec)\n",
"INFO:tensorflow:global_step/sec: 1.53778\n",
"INFO:tensorflow:train_accuracy = 0.3604061, train_loss = 34.712807 (0.650 sec)\n",
"INFO:tensorflow:loss = 34.712807, step = 257 (0.651 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64734\n",
"INFO:tensorflow:train_accuracy = 0.35135135, train_loss = 30.506767 (0.607 sec)\n",
"INFO:tensorflow:loss = 30.506767, step = 258 (0.607 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88136\n",
"INFO:tensorflow:train_accuracy = 0.39030612, train_loss = 28.8061 (0.531 sec)\n",
"INFO:tensorflow:loss = 28.8061, step = 259 (0.531 sec)\n",
"INFO:tensorflow:global_step/sec: 1.52631\n",
"INFO:tensorflow:train_accuracy = 0.3320659, train_loss = 35.30445 (0.655 sec)\n",
"INFO:tensorflow:loss = 35.30445, step = 260 (0.655 sec)\n",
"INFO:tensorflow:global_step/sec: 1.9953\n",
"INFO:tensorflow:train_accuracy = 0.36129034, train_loss = 27.489025 (0.501 sec)\n",
"INFO:tensorflow:loss = 27.489025, step = 261 (0.501 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74591\n",
"INFO:tensorflow:train_accuracy = 0.36479592, train_loss = 27.586046 (0.572 sec)\n",
"INFO:tensorflow:loss = 27.586046, step = 262 (0.572 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65629\n",
"INFO:tensorflow:train_accuracy = 0.44542032, train_loss = 32.163177 (0.604 sec)\n",
"INFO:tensorflow:loss = 32.163177, step = 263 (0.604 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86774\n",
"INFO:tensorflow:train_accuracy = 0.40632913, train_loss = 35.837086 (0.536 sec)\n",
"INFO:tensorflow:loss = 35.837086, step = 264 (0.536 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73753\n",
"INFO:tensorflow:train_accuracy = 0.4108911, train_loss = 30.432001 (0.576 sec)\n",
"INFO:tensorflow:loss = 30.432001, step = 265 (0.575 sec)\n",
"INFO:tensorflow:global_step/sec: 1.54804\n",
"INFO:tensorflow:train_accuracy = 0.28192163, train_loss = 37.065895 (0.646 sec)\n",
"INFO:tensorflow:loss = 37.065895, step = 266 (0.646 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69338\n",
"INFO:tensorflow:train_accuracy = 0.3587385, train_loss = 25.964443 (0.590 sec)\n",
"INFO:tensorflow:loss = 25.964443, step = 267 (0.590 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89636\n",
"INFO:tensorflow:train_accuracy = 0.34709677, train_loss = 32.253067 (0.527 sec)\n",
"INFO:tensorflow:loss = 32.253067, step = 268 (0.527 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72029\n",
"INFO:tensorflow:train_accuracy = 0.33333334, train_loss = 28.457481 (0.582 sec)\n",
"INFO:tensorflow:loss = 28.457481, step = 269 (0.582 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66782\n",
"INFO:tensorflow:train_accuracy = 0.35751295, train_loss = 31.574627 (0.600 sec)\n",
"INFO:tensorflow:loss = 31.574627, step = 270 (0.599 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61031\n",
"INFO:tensorflow:train_accuracy = 0.45420793, train_loss = 32.956604 (0.621 sec)\n",
"INFO:tensorflow:loss = 32.956604, step = 271 (0.621 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82406\n",
"INFO:tensorflow:train_accuracy = 0.39572865, train_loss = 32.288895 (0.549 sec)\n",
"INFO:tensorflow:loss = 32.288895, step = 272 (0.549 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69031\n",
"INFO:tensorflow:train_accuracy = 0.46857923, train_loss = 24.622152 (0.591 sec)\n",
"INFO:tensorflow:loss = 24.622152, step = 273 (0.591 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89458\n",
"INFO:tensorflow:train_accuracy = 0.35479632, train_loss = 26.510595 (0.528 sec)\n",
"INFO:tensorflow:loss = 26.510595, step = 274 (0.528 sec)\n",
"INFO:tensorflow:global_step/sec: 4.16001\n",
"INFO:tensorflow:train_accuracy = 0.36138615, train_loss = 23.861013 (0.240 sec)\n",
"INFO:tensorflow:loss = 23.861013, step = 275 (0.240 sec)\n",
"INFO:tensorflow:global_step/sec: 1.52784\n",
"INFO:tensorflow:train_accuracy = 0.36950904, train_loss = 28.181776 (0.655 sec)\n",
"INFO:tensorflow:loss = 28.181776, step = 276 (0.655 sec)\n",
"INFO:tensorflow:global_step/sec: 1.42884\n",
"INFO:tensorflow:train_accuracy = 0.36239102, train_loss = 36.481026 (0.700 sec)\n",
"INFO:tensorflow:loss = 36.481026, step = 277 (0.700 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66312\n",
"INFO:tensorflow:train_accuracy = 0.28337532, train_loss = 35.560234 (0.601 sec)\n",
"INFO:tensorflow:loss = 35.560234, step = 278 (0.602 sec)\n",
"INFO:tensorflow:global_step/sec: 1.79778\n",
"INFO:tensorflow:train_accuracy = 0.387985, train_loss = 29.7589 (0.556 sec)\n",
"INFO:tensorflow:loss = 29.7589, step = 279 (0.556 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88358\n",
"INFO:tensorflow:train_accuracy = 0.46084723, train_loss = 28.123295 (0.531 sec)\n",
"INFO:tensorflow:loss = 28.123295, step = 280 (0.531 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64112\n",
"INFO:tensorflow:train_accuracy = 0.46692607, train_loss = 31.535358 (0.609 sec)\n",
"INFO:tensorflow:loss = 31.535358, step = 281 (0.610 sec)\n",
"INFO:tensorflow:global_step/sec: 1.68917\n",
"INFO:tensorflow:train_accuracy = 0.49608356, train_loss = 26.253849 (0.592 sec)\n",
"INFO:tensorflow:loss = 26.253849, step = 282 (0.592 sec)\n",
"INFO:tensorflow:global_step/sec: 2.09021\n",
"INFO:tensorflow:train_accuracy = 0.39021164, train_loss = 25.513786 (0.479 sec)\n",
"INFO:tensorflow:loss = 25.513786, step = 283 (0.479 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62835\n",
"INFO:tensorflow:train_accuracy = 0.36178344, train_loss = 28.07068 (0.614 sec)\n",
"INFO:tensorflow:loss = 28.07068, step = 284 (0.614 sec)\n",
"INFO:tensorflow:global_step/sec: 1.44622\n",
"INFO:tensorflow:train_accuracy = 0.39328858, train_loss = 28.481018 (0.691 sec)\n",
"INFO:tensorflow:loss = 28.481018, step = 285 (0.691 sec)\n",
"INFO:tensorflow:global_step/sec: 1.93583\n",
"INFO:tensorflow:train_accuracy = 0.37715805, train_loss = 24.430786 (0.517 sec)\n",
"INFO:tensorflow:loss = 24.430786, step = 286 (0.517 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64758\n",
"INFO:tensorflow:train_accuracy = 0.39333335, train_loss = 24.877079 (0.607 sec)\n",
"INFO:tensorflow:loss = 24.877079, step = 287 (0.606 sec)\n",
"INFO:tensorflow:global_step/sec: 1.81981\n",
"INFO:tensorflow:train_accuracy = 0.40279543, train_loss = 27.373955 (0.550 sec)\n",
"INFO:tensorflow:loss = 27.373955, step = 288 (0.550 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90762\n",
"INFO:tensorflow:train_accuracy = 0.364, train_loss = 26.825285 (0.524 sec)\n",
"INFO:tensorflow:loss = 26.825285, step = 289 (0.524 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88558\n",
"INFO:tensorflow:train_accuracy = 0.3615894, train_loss = 25.449526 (0.530 sec)\n",
"INFO:tensorflow:loss = 25.449526, step = 290 (0.530 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88657\n",
"INFO:tensorflow:train_accuracy = 0.37579617, train_loss = 27.94577 (0.530 sec)\n",
"INFO:tensorflow:loss = 27.94577, step = 291 (0.531 sec)\n",
"INFO:tensorflow:global_step/sec: 1.79311\n",
"INFO:tensorflow:train_accuracy = 0.29123712, train_loss = 30.154083 (0.557 sec)\n",
"INFO:tensorflow:loss = 30.154083, step = 292 (0.557 sec)\n",
"INFO:tensorflow:global_step/sec: 2.04027\n",
"INFO:tensorflow:train_accuracy = 0.42801556, train_loss = 28.020037 (0.490 sec)\n",
"INFO:tensorflow:loss = 28.020037, step = 293 (0.491 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80691\n",
"INFO:tensorflow:train_accuracy = 0.35564303, train_loss = 26.548489 (0.553 sec)\n",
"INFO:tensorflow:loss = 26.548489, step = 294 (0.553 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8423\n",
"INFO:tensorflow:train_accuracy = 0.5013477, train_loss = 24.833122 (0.543 sec)\n",
"INFO:tensorflow:loss = 24.833122, step = 295 (0.543 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61861\n",
"INFO:tensorflow:train_accuracy = 0.45, train_loss = 35.588493 (0.618 sec)\n",
"INFO:tensorflow:loss = 35.588493, step = 296 (0.618 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86872\n",
"INFO:tensorflow:train_accuracy = 0.5099867, train_loss = 25.103636 (0.535 sec)\n",
"INFO:tensorflow:loss = 25.103636, step = 297 (0.535 sec)\n",
"INFO:tensorflow:global_step/sec: 1.84721\n",
"INFO:tensorflow:train_accuracy = 0.47388536, train_loss = 29.936462 (0.541 sec)\n",
"INFO:tensorflow:loss = 29.936462, step = 298 (0.542 sec)\n",
"INFO:tensorflow:global_step/sec: 2.00271\n",
"INFO:tensorflow:train_accuracy = 0.47251308, train_loss = 25.765974 (0.500 sec)\n",
"INFO:tensorflow:loss = 25.765974, step = 299 (0.500 sec)\n",
"INFO:tensorflow:global_step/sec: 1.76846\n",
"INFO:tensorflow:train_accuracy = 0.4898236, train_loss = 26.566757 (0.566 sec)\n",
"INFO:tensorflow:loss = 26.566757, step = 300 (0.566 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77972\n",
"INFO:tensorflow:train_accuracy = 0.49425286, train_loss = 27.697672 (0.561 sec)\n",
"INFO:tensorflow:loss = 27.697672, step = 301 (0.561 sec)\n",
"INFO:tensorflow:global_step/sec: 1.94777\n",
"INFO:tensorflow:train_accuracy = 0.4616402, train_loss = 25.210155 (0.513 sec)\n",
"INFO:tensorflow:loss = 25.210155, step = 302 (0.513 sec)\n",
"INFO:tensorflow:global_step/sec: 1.5484\n",
"INFO:tensorflow:train_accuracy = 0.4435995, train_loss = 37.044415 (0.646 sec)\n",
"INFO:tensorflow:loss = 37.044415, step = 303 (0.646 sec)\n",
"INFO:tensorflow:global_step/sec: 1.96475\n",
"INFO:tensorflow:train_accuracy = 0.45801526, train_loss = 28.910404 (0.509 sec)\n",
"INFO:tensorflow:loss = 28.910404, step = 304 (0.509 sec)\n",
"INFO:tensorflow:global_step/sec: 2.04004\n",
"INFO:tensorflow:train_accuracy = 0.44416243, train_loss = 29.92569 (0.490 sec)\n",
"INFO:tensorflow:loss = 29.92569, step = 305 (0.490 sec)\n",
"INFO:tensorflow:global_step/sec: 1.52896\n",
"INFO:tensorflow:train_accuracy = 0.44233206, train_loss = 32.637974 (0.655 sec)\n",
"INFO:tensorflow:loss = 32.637974, step = 306 (0.655 sec)\n",
"INFO:tensorflow:global_step/sec: 1.96769\n",
"INFO:tensorflow:train_accuracy = 0.46460748, train_loss = 26.920086 (0.508 sec)\n",
"INFO:tensorflow:loss = 26.920086, step = 307 (0.507 sec)\n",
"INFO:tensorflow:global_step/sec: 2.06877\n",
"INFO:tensorflow:train_accuracy = 0.46989527, train_loss = 26.642605 (0.483 sec)\n",
"INFO:tensorflow:loss = 26.642605, step = 308 (0.484 sec)\n",
"INFO:tensorflow:global_step/sec: 1.6303\n",
"INFO:tensorflow:train_accuracy = 0.41430333, train_loss = 29.293068 (0.614 sec)\n",
"INFO:tensorflow:loss = 29.293068, step = 309 (0.614 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65586\n",
"INFO:tensorflow:train_accuracy = 0.48358586, train_loss = 29.308533 (0.604 sec)\n",
"INFO:tensorflow:loss = 29.308533, step = 310 (0.604 sec)\n",
"INFO:tensorflow:global_step/sec: 1.54163\n",
"INFO:tensorflow:train_accuracy = 0.4224138, train_loss = 32.32497 (0.648 sec)\n",
"INFO:tensorflow:loss = 32.32497, step = 311 (0.648 sec)\n",
"INFO:tensorflow:global_step/sec: 1.98295\n",
"INFO:tensorflow:train_accuracy = 0.44025975, train_loss = 26.00887 (0.504 sec)\n",
"INFO:tensorflow:loss = 26.00887, step = 312 (0.504 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86642\n",
"INFO:tensorflow:train_accuracy = 0.45032257, train_loss = 27.37177 (0.536 sec)\n",
"INFO:tensorflow:loss = 27.37177, step = 313 (0.535 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72383\n",
"INFO:tensorflow:train_accuracy = 0.46944082, train_loss = 25.500511 (0.580 sec)\n",
"INFO:tensorflow:loss = 25.500511, step = 314 (0.580 sec)\n",
"INFO:tensorflow:global_step/sec: 1.99755\n",
"INFO:tensorflow:train_accuracy = 0.3455696, train_loss = 29.397465 (0.501 sec)\n",
"INFO:tensorflow:loss = 29.397465, step = 315 (0.501 sec)\n",
"INFO:tensorflow:global_step/sec: 1.70932\n",
"INFO:tensorflow:train_accuracy = 0.4552333, train_loss = 30.281363 (0.585 sec)\n",
"INFO:tensorflow:loss = 30.281363, step = 316 (0.585 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7393\n",
"INFO:tensorflow:train_accuracy = 0.5046113, train_loss = 26.507818 (0.575 sec)\n",
"INFO:tensorflow:loss = 26.507818, step = 317 (0.575 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88963\n",
"INFO:tensorflow:train_accuracy = 0.48236775, train_loss = 28.591263 (0.529 sec)\n",
"INFO:tensorflow:loss = 28.591263, step = 318 (0.529 sec)\n",
"INFO:tensorflow:global_step/sec: 1.95754\n",
"INFO:tensorflow:train_accuracy = 0.49290323, train_loss = 26.254997 (0.511 sec)\n",
"INFO:tensorflow:loss = 26.254997, step = 319 (0.511 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91259\n",
"INFO:tensorflow:train_accuracy = 0.5020243, train_loss = 22.545403 (0.523 sec)\n",
"INFO:tensorflow:loss = 22.545403, step = 320 (0.523 sec)\n",
"INFO:tensorflow:global_step/sec: 5.69239\n",
"INFO:tensorflow:train_accuracy = 0.51960784, train_loss = 21.384544 (0.175 sec)\n",
"INFO:tensorflow:loss = 21.384544, step = 321 (0.175 sec)\n",
"INFO:tensorflow:global_step/sec: 1.70913\n",
"INFO:tensorflow:train_accuracy = 0.47035572, train_loss = 32.11174 (0.585 sec)\n",
"INFO:tensorflow:loss = 32.11174, step = 322 (0.585 sec)\n",
"INFO:tensorflow:global_step/sec: 1.42643\n",
"INFO:tensorflow:train_accuracy = 0.44625, train_loss = 37.92308 (0.701 sec)\n",
"INFO:tensorflow:loss = 37.92308, step = 323 (0.701 sec)\n",
"INFO:tensorflow:global_step/sec: 1.83751\n",
"INFO:tensorflow:train_accuracy = 0.4757764, train_loss = 28.940554 (0.545 sec)\n",
"INFO:tensorflow:loss = 28.940554, step = 324 (0.545 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80044\n",
"INFO:tensorflow:train_accuracy = 0.4534005, train_loss = 28.306011 (0.555 sec)\n",
"INFO:tensorflow:loss = 28.306011, step = 325 (0.555 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63036\n",
"INFO:tensorflow:train_accuracy = 0.4786546, train_loss = 26.277744 (0.613 sec)\n",
"INFO:tensorflow:loss = 26.277744, step = 326 (0.613 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69798\n",
"INFO:tensorflow:train_accuracy = 0.47381866, train_loss = 26.768934 (0.589 sec)\n",
"INFO:tensorflow:loss = 26.768934, step = 327 (0.589 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73245\n",
"INFO:tensorflow:train_accuracy = 0.50797874, train_loss = 28.927519 (0.577 sec)\n",
"INFO:tensorflow:loss = 28.927519, step = 328 (0.578 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69081\n",
"INFO:tensorflow:train_accuracy = 0.5097784, train_loss = 26.517214 (0.592 sec)\n",
"INFO:tensorflow:loss = 26.517214, step = 329 (0.592 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62745\n",
"INFO:tensorflow:train_accuracy = 0.47532466, train_loss = 25.827774 (0.614 sec)\n",
"INFO:tensorflow:loss = 25.827774, step = 330 (0.614 sec)\n",
"INFO:tensorflow:global_step/sec: 1.98732\n",
"INFO:tensorflow:train_accuracy = 0.48472777, train_loss = 26.021025 (0.503 sec)\n",
"INFO:tensorflow:loss = 26.021025, step = 331 (0.503 sec)\n",
"INFO:tensorflow:global_step/sec: 1.45338\n",
"INFO:tensorflow:train_accuracy = 0.47619048, train_loss = 28.046333 (0.688 sec)\n",
"INFO:tensorflow:loss = 28.046333, step = 332 (0.688 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89903\n",
"INFO:tensorflow:train_accuracy = 0.47686374, train_loss = 27.294468 (0.526 sec)\n",
"INFO:tensorflow:loss = 27.294468, step = 333 (0.526 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90765\n",
"INFO:tensorflow:train_accuracy = 0.48840207, train_loss = 26.042688 (0.524 sec)\n",
"INFO:tensorflow:loss = 26.042688, step = 334 (0.525 sec)\n",
"INFO:tensorflow:global_step/sec: 1.97467\n",
"INFO:tensorflow:train_accuracy = 0.47011954, train_loss = 25.19096 (0.507 sec)\n",
"INFO:tensorflow:loss = 25.19096, step = 335 (0.507 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64299\n",
"INFO:tensorflow:train_accuracy = 0.48883048, train_loss = 26.67872 (0.608 sec)\n",
"INFO:tensorflow:loss = 26.67872, step = 336 (0.608 sec)\n",
"INFO:tensorflow:global_step/sec: 1.836\n",
"INFO:tensorflow:train_accuracy = 0.45918366, train_loss = 27.486906 (0.545 sec)\n",
"INFO:tensorflow:loss = 27.486906, step = 337 (0.545 sec)\n",
"INFO:tensorflow:global_step/sec: 1.93956\n",
"INFO:tensorflow:train_accuracy = 0.47424042, train_loss = 25.39995 (0.516 sec)\n",
"INFO:tensorflow:loss = 25.39995, step = 338 (0.516 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87626\n",
"INFO:tensorflow:train_accuracy = 0.49559748, train_loss = 29.533787 (0.533 sec)\n",
"INFO:tensorflow:loss = 29.533787, step = 339 (0.533 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8538\n",
"INFO:tensorflow:train_accuracy = 0.50406504, train_loss = 22.796747 (0.539 sec)\n",
"INFO:tensorflow:loss = 22.796747, step = 340 (0.539 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82943\n",
"INFO:tensorflow:train_accuracy = 0.48235294, train_loss = 28.653038 (0.546 sec)\n",
"INFO:tensorflow:loss = 28.653038, step = 341 (0.546 sec)\n",
"INFO:tensorflow:global_step/sec: 1.58142\n",
"INFO:tensorflow:train_accuracy = 0.4602564, train_loss = 30.651745 (0.633 sec)\n",
"INFO:tensorflow:loss = 30.651745, step = 342 (0.633 sec)\n",
"INFO:tensorflow:global_step/sec: 2.00013\n",
"INFO:tensorflow:train_accuracy = 0.5079576, train_loss = 23.703407 (0.500 sec)\n",
"INFO:tensorflow:loss = 23.703407, step = 343 (0.500 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82419\n",
"INFO:tensorflow:train_accuracy = 0.50325096, train_loss = 24.884956 (0.548 sec)\n",
"INFO:tensorflow:loss = 24.884956, step = 344 (0.547 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86094\n",
"INFO:tensorflow:train_accuracy = 0.4654498, train_loss = 25.83979 (0.538 sec)\n",
"INFO:tensorflow:loss = 25.83979, step = 345 (0.538 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88649\n",
"INFO:tensorflow:train_accuracy = 0.5006553, train_loss = 24.278006 (0.530 sec)\n",
"INFO:tensorflow:loss = 24.278006, step = 346 (0.530 sec)\n",
"INFO:tensorflow:global_step/sec: 1.76439\n",
"INFO:tensorflow:train_accuracy = 0.500664, train_loss = 25.942574 (0.567 sec)\n",
"INFO:tensorflow:loss = 25.942574, step = 347 (0.567 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62249\n",
"INFO:tensorflow:train_accuracy = 0.49039692, train_loss = 26.203129 (0.616 sec)\n",
"INFO:tensorflow:loss = 26.203129, step = 348 (0.616 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74333\n",
"INFO:tensorflow:train_accuracy = 0.4474359, train_loss = 30.295635 (0.573 sec)\n",
"INFO:tensorflow:loss = 30.295635, step = 349 (0.573 sec)\n",
"INFO:tensorflow:global_step/sec: 1.5293\n",
"INFO:tensorflow:train_accuracy = 0.45700246, train_loss = 36.840958 (0.654 sec)\n",
"INFO:tensorflow:loss = 36.840958, step = 350 (0.655 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69309\n",
"INFO:tensorflow:train_accuracy = 0.49165598, train_loss = 26.732471 (0.591 sec)\n",
"INFO:tensorflow:loss = 26.732471, step = 351 (0.591 sec)\n",
"INFO:tensorflow:global_step/sec: 1.49011\n",
"INFO:tensorflow:train_accuracy = 0.4744898, train_loss = 31.166584 (0.671 sec)\n",
"INFO:tensorflow:loss = 31.166584, step = 352 (0.672 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62692\n",
"INFO:tensorflow:train_accuracy = 0.4866582, train_loss = 26.602047 (0.614 sec)\n",
"INFO:tensorflow:loss = 26.602047, step = 353 (0.614 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72826\n",
"INFO:tensorflow:train_accuracy = 0.5045161, train_loss = 29.540314 (0.578 sec)\n",
"INFO:tensorflow:loss = 29.540314, step = 354 (0.578 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75612\n",
"INFO:tensorflow:train_accuracy = 0.46163684, train_loss = 28.468475 (0.569 sec)\n",
"INFO:tensorflow:loss = 28.468475, step = 355 (0.569 sec)\n",
"INFO:tensorflow:global_step/sec: 1.92884\n",
"INFO:tensorflow:train_accuracy = 0.50732356, train_loss = 22.810867 (0.519 sec)\n",
"INFO:tensorflow:loss = 22.810867, step = 356 (0.519 sec)\n",
"INFO:tensorflow:global_step/sec: 1.54877\n",
"INFO:tensorflow:train_accuracy = 0.4547619, train_loss = 34.23915 (0.646 sec)\n",
"INFO:tensorflow:loss = 34.23915, step = 357 (0.645 sec)\n",
"INFO:tensorflow:global_step/sec: 1.84032\n",
"INFO:tensorflow:train_accuracy = 0.49738905, train_loss = 37.912354 (0.544 sec)\n",
"INFO:tensorflow:loss = 37.912354, step = 358 (0.544 sec)\n",
"INFO:tensorflow:global_step/sec: 1.9038\n",
"INFO:tensorflow:train_accuracy = 0.49281046, train_loss = 32.12382 (0.525 sec)\n",
"INFO:tensorflow:loss = 32.12382, step = 359 (0.525 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89727\n",
"INFO:tensorflow:train_accuracy = 0.5070243, train_loss = 27.803692 (0.527 sec)\n",
"INFO:tensorflow:loss = 27.803692, step = 360 (0.527 sec)\n",
"INFO:tensorflow:global_step/sec: 2.05864\n",
"INFO:tensorflow:train_accuracy = 0.48992443, train_loss = 35.985584 (0.486 sec)\n",
"INFO:tensorflow:loss = 35.985584, step = 361 (0.486 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74664\n",
"INFO:tensorflow:train_accuracy = 0.47637796, train_loss = 35.16639 (0.573 sec)\n",
"INFO:tensorflow:loss = 35.16639, step = 362 (0.573 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69213\n",
"INFO:tensorflow:train_accuracy = 0.44389027, train_loss = 35.99266 (0.591 sec)\n",
"INFO:tensorflow:loss = 35.99266, step = 363 (0.591 sec)\n",
"INFO:tensorflow:global_step/sec: 1.94956\n",
"INFO:tensorflow:train_accuracy = 0.47900262, train_loss = 32.394775 (0.513 sec)\n",
"INFO:tensorflow:loss = 32.394775, step = 364 (0.513 sec)\n",
"INFO:tensorflow:global_step/sec: 1.84088\n",
"INFO:tensorflow:train_accuracy = 0.4784876, train_loss = 33.95408 (0.544 sec)\n",
"INFO:tensorflow:loss = 33.95408, step = 365 (0.544 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86874\n",
"INFO:tensorflow:train_accuracy = 0.40629095, train_loss = 31.711056 (0.535 sec)\n",
"INFO:tensorflow:loss = 31.711056, step = 366 (0.535 sec)\n",
"INFO:tensorflow:global_step/sec: 4.09304\n",
"INFO:tensorflow:train_accuracy = 0.4232558, train_loss = 36.376175 (0.244 sec)\n",
"INFO:tensorflow:loss = 36.376175, step = 367 (0.244 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66335\n",
"INFO:tensorflow:train_accuracy = 0.41229194, train_loss = 36.46441 (0.601 sec)\n",
"INFO:tensorflow:loss = 36.46441, step = 368 (0.601 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64908\n",
"INFO:tensorflow:train_accuracy = 0.46023467, train_loss = 30.600323 (0.606 sec)\n",
"INFO:tensorflow:loss = 30.600323, step = 369 (0.606 sec)\n",
"INFO:tensorflow:global_step/sec: 1.41997\n",
"INFO:tensorflow:train_accuracy = 0.40197286, train_loss = 42.17343 (0.704 sec)\n",
"INFO:tensorflow:loss = 42.17343, step = 370 (0.704 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74036\n",
"INFO:tensorflow:train_accuracy = 0.40300375, train_loss = 37.715996 (0.575 sec)\n",
"INFO:tensorflow:loss = 37.715996, step = 371 (0.575 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89623\n",
"INFO:tensorflow:train_accuracy = 0.32818532, train_loss = 45.74591 (0.527 sec)\n",
"INFO:tensorflow:loss = 45.74591, step = 372 (0.527 sec)\n",
"INFO:tensorflow:global_step/sec: 1.70167\n",
"INFO:tensorflow:train_accuracy = 0.29551452, train_loss = 33.569115 (0.588 sec)\n",
"INFO:tensorflow:loss = 33.569115, step = 373 (0.588 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63822\n",
"INFO:tensorflow:train_accuracy = 0.31806615, train_loss = 36.659058 (0.611 sec)\n",
"INFO:tensorflow:loss = 36.659058, step = 374 (0.611 sec)\n",
"INFO:tensorflow:global_step/sec: 2.00118\n",
"INFO:tensorflow:train_accuracy = 0.30820104, train_loss = 29.066776 (0.499 sec)\n",
"INFO:tensorflow:loss = 29.066776, step = 375 (0.499 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63956\n",
"INFO:tensorflow:train_accuracy = 0.34895834, train_loss = 30.917545 (0.610 sec)\n",
"INFO:tensorflow:loss = 30.917545, step = 376 (0.610 sec)\n",
"INFO:tensorflow:global_step/sec: 1.4545\n",
"INFO:tensorflow:train_accuracy = 0.39686683, train_loss = 35.733757 (0.687 sec)\n",
"INFO:tensorflow:loss = 35.733757, step = 377 (0.687 sec)\n",
"INFO:tensorflow:global_step/sec: 2.08324\n",
"INFO:tensorflow:train_accuracy = 0.4315515, train_loss = 29.345798 (0.480 sec)\n",
"INFO:tensorflow:loss = 29.345798, step = 378 (0.480 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62978\n",
"INFO:tensorflow:train_accuracy = 0.3617571, train_loss = 32.060646 (0.614 sec)\n",
"INFO:tensorflow:loss = 32.060646, step = 379 (0.614 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89767\n",
"INFO:tensorflow:train_accuracy = 0.38881665, train_loss = 29.178951 (0.527 sec)\n",
"INFO:tensorflow:loss = 29.178951, step = 380 (0.527 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86942\n",
"INFO:tensorflow:train_accuracy = 0.4063745, train_loss = 27.658041 (0.535 sec)\n",
"INFO:tensorflow:loss = 27.658041, step = 381 (0.534 sec)\n",
"INFO:tensorflow:global_step/sec: 1.98106\n",
"INFO:tensorflow:train_accuracy = 0.45794392, train_loss = 27.514187 (0.505 sec)\n",
"INFO:tensorflow:loss = 27.514187, step = 382 (0.506 sec)\n",
"INFO:tensorflow:global_step/sec: 1.78529\n",
"INFO:tensorflow:train_accuracy = 0.4450194, train_loss = 29.60213 (0.560 sec)\n",
"INFO:tensorflow:loss = 29.60213, step = 383 (0.559 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86033\n",
"INFO:tensorflow:train_accuracy = 0.4502551, train_loss = 28.038002 (0.538 sec)\n",
"INFO:tensorflow:loss = 28.038002, step = 384 (0.539 sec)\n",
"INFO:tensorflow:global_step/sec: 2.01412\n",
"INFO:tensorflow:train_accuracy = 0.47643313, train_loss = 29.155914 (0.496 sec)\n",
"INFO:tensorflow:loss = 29.155914, step = 385 (0.496 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91093\n",
"INFO:tensorflow:train_accuracy = 0.5154778, train_loss = 25.080322 (0.523 sec)\n",
"INFO:tensorflow:loss = 25.080322, step = 386 (0.523 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80567\n",
"INFO:tensorflow:train_accuracy = 0.505291, train_loss = 30.942326 (0.554 sec)\n",
"INFO:tensorflow:loss = 30.942326, step = 387 (0.554 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87053\n",
"INFO:tensorflow:train_accuracy = 0.52489907, train_loss = 23.914503 (0.535 sec)\n",
"INFO:tensorflow:loss = 23.914503, step = 388 (0.535 sec)\n",
"INFO:tensorflow:global_step/sec: 1.60822\n",
"INFO:tensorflow:train_accuracy = 0.49622166, train_loss = 29.294128 (0.621 sec)\n",
"INFO:tensorflow:loss = 29.294128, step = 389 (0.621 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85431\n",
"INFO:tensorflow:train_accuracy = 0.46575344, train_loss = 31.312845 (0.540 sec)\n",
"INFO:tensorflow:loss = 31.312845, step = 390 (0.540 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75239\n",
"INFO:tensorflow:train_accuracy = 0.500657, train_loss = 28.611073 (0.571 sec)\n",
"INFO:tensorflow:loss = 28.611073, step = 391 (0.570 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89632\n",
"INFO:tensorflow:train_accuracy = 0.5472973, train_loss = 22.707668 (0.528 sec)\n",
"INFO:tensorflow:loss = 22.707668, step = 392 (0.528 sec)\n",
"INFO:tensorflow:global_step/sec: 1.83605\n",
"INFO:tensorflow:train_accuracy = 0.48425198, train_loss = 25.295143 (0.544 sec)\n",
"INFO:tensorflow:loss = 25.295143, step = 393 (0.544 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73506\n",
"INFO:tensorflow:train_accuracy = 0.4934037, train_loss = 25.991133 (0.576 sec)\n",
"INFO:tensorflow:loss = 25.991133, step = 394 (0.577 sec)\n",
"INFO:tensorflow:global_step/sec: 1.52749\n",
"INFO:tensorflow:train_accuracy = 0.48692402, train_loss = 38.746124 (0.655 sec)\n",
"INFO:tensorflow:loss = 38.746124, step = 395 (0.655 sec)\n",
"INFO:tensorflow:global_step/sec: 1.79087\n",
"INFO:tensorflow:train_accuracy = 0.4897436, train_loss = 28.773373 (0.558 sec)\n",
"INFO:tensorflow:loss = 28.773373, step = 396 (0.558 sec)\n",
"INFO:tensorflow:global_step/sec: 1.98969\n",
"INFO:tensorflow:train_accuracy = 0.5192563, train_loss = 26.096806 (0.503 sec)\n",
"INFO:tensorflow:loss = 26.096806, step = 397 (0.502 sec)\n",
"INFO:tensorflow:global_step/sec: 1.94987\n",
"INFO:tensorflow:train_accuracy = 0.51007557, train_loss = 27.813423 (0.513 sec)\n",
"INFO:tensorflow:loss = 27.813423, step = 398 (0.513 sec)\n",
"INFO:tensorflow:Saving checkpoints for 400 into asr-quartznet/model.ckpt.\n",
"INFO:tensorflow:Calling model_fn.\n",
"INFO:tensorflow:Done calling model_fn.\n",
"INFO:tensorflow:Starting evaluation at 2020-10-25T18:58:43Z\n",
"INFO:tensorflow:Graph was finalized.\n",
"INFO:tensorflow:Restoring parameters from asr-quartznet/model.ckpt-400\n",
"INFO:tensorflow:Running local_init_op.\n",
"INFO:tensorflow:Done running local_init_op.\n",
"INFO:tensorflow:Evaluation [1/10]\n",
"INFO:tensorflow:Evaluation [2/10]\n",
"INFO:tensorflow:Evaluation [3/10]\n",
"INFO:tensorflow:Evaluation [4/10]\n",
"INFO:tensorflow:Evaluation [5/10]\n",
"INFO:tensorflow:Evaluation [6/10]\n",
"INFO:tensorflow:Evaluation [7/10]\n",
"INFO:tensorflow:Evaluation [8/10]\n",
"INFO:tensorflow:Evaluation [9/10]\n",
"INFO:tensorflow:Evaluation [10/10]\n",
"INFO:tensorflow:Finished evaluation at 2020-10-25-18:58:52\n",
"INFO:tensorflow:Saving dict for global step 400: WER = 1.0, accuracy = 0.4815951, global_step = 400, loss = 25.727505\n",
"INFO:tensorflow:Saving 'checkpoint_path' summary for global step 400: asr-quartznet/model.ckpt-400\n",
"INFO:tensorflow:global_step/sec: 0.0577697\n",
"INFO:tensorflow:train_accuracy = 0.53149605, train_loss = 25.116385 (17.310 sec)\n",
"INFO:tensorflow:loss = 25.116385, step = 399 (17.310 sec)\n",
"INFO:tensorflow:global_step/sec: 1.4284\n",
"INFO:tensorflow:train_accuracy = 0.48399016, train_loss = 32.16086 (0.701 sec)\n",
"INFO:tensorflow:loss = 32.16086, step = 400 (0.701 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85709\n",
"INFO:tensorflow:train_accuracy = 0.475032, train_loss = 29.04078 (0.538 sec)\n",
"INFO:tensorflow:loss = 29.04078, step = 401 (0.538 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72061\n",
"INFO:tensorflow:train_accuracy = 0.47013977, train_loss = 29.42472 (0.581 sec)\n",
"INFO:tensorflow:loss = 29.42472, step = 402 (0.581 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69614\n",
"INFO:tensorflow:train_accuracy = 0.47459728, train_loss = 30.10174 (0.589 sec)\n",
"INFO:tensorflow:loss = 30.10174, step = 403 (0.589 sec)\n",
"INFO:tensorflow:global_step/sec: 1.67361\n",
"INFO:tensorflow:train_accuracy = 0.50383633, train_loss = 25.784248 (0.598 sec)\n",
"INFO:tensorflow:loss = 25.784248, step = 404 (0.598 sec)\n",
"INFO:tensorflow:global_step/sec: 1.56002\n",
"INFO:tensorflow:train_accuracy = 0.49551857, train_loss = 29.665733 (0.641 sec)\n",
"INFO:tensorflow:loss = 29.665733, step = 405 (0.641 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89875\n",
"INFO:tensorflow:train_accuracy = 0.4974293, train_loss = 26.556713 (0.526 sec)\n",
"INFO:tensorflow:loss = 26.556713, step = 406 (0.527 sec)\n",
"INFO:tensorflow:global_step/sec: 1.98731\n",
"INFO:tensorflow:train_accuracy = 0.5006468, train_loss = 28.414324 (0.503 sec)\n",
"INFO:tensorflow:loss = 28.414324, step = 407 (0.503 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72623\n",
"INFO:tensorflow:train_accuracy = 0.4925, train_loss = 29.401793 (0.579 sec)\n",
"INFO:tensorflow:loss = 29.401793, step = 408 (0.579 sec)\n",
"INFO:tensorflow:global_step/sec: 1.78302\n",
"INFO:tensorflow:train_accuracy = 0.5232558, train_loss = 25.468164 (0.561 sec)\n",
"INFO:tensorflow:loss = 25.468164, step = 409 (0.561 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74695\n",
"INFO:tensorflow:train_accuracy = 0.51861364, train_loss = 27.467693 (0.573 sec)\n",
"INFO:tensorflow:loss = 27.467693, step = 410 (0.572 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82632\n",
"INFO:tensorflow:train_accuracy = 0.5123537, train_loss = 25.783855 (0.548 sec)\n",
"INFO:tensorflow:loss = 25.783855, step = 411 (0.548 sec)\n",
"INFO:tensorflow:global_step/sec: 1.93926\n",
"INFO:tensorflow:train_accuracy = 0.54508746, train_loss = 22.686821 (0.516 sec)\n",
"INFO:tensorflow:loss = 22.686821, step = 412 (0.515 sec)\n",
"INFO:tensorflow:global_step/sec: 5.60299\n",
"INFO:tensorflow:train_accuracy = 0.43055555, train_loss = 26.183275 (0.179 sec)\n",
"INFO:tensorflow:loss = 26.183275, step = 413 (0.179 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64514\n",
"INFO:tensorflow:train_accuracy = 0.5135484, train_loss = 26.06844 (0.608 sec)\n",
"INFO:tensorflow:loss = 26.06844, step = 414 (0.608 sec)\n",
"INFO:tensorflow:global_step/sec: 1.44528\n",
"INFO:tensorflow:train_accuracy = 0.5068836, train_loss = 28.857237 (0.692 sec)\n",
"INFO:tensorflow:loss = 28.857237, step = 415 (0.692 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64673\n",
"INFO:tensorflow:train_accuracy = 0.4778761, train_loss = 29.197338 (0.607 sec)\n",
"INFO:tensorflow:loss = 29.197338, step = 416 (0.607 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7929\n",
"INFO:tensorflow:train_accuracy = 0.49127182, train_loss = 29.227911 (0.557 sec)\n",
"INFO:tensorflow:loss = 29.227911, step = 417 (0.557 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61995\n",
"INFO:tensorflow:train_accuracy = 0.50319284, train_loss = 29.551336 (0.618 sec)\n",
"INFO:tensorflow:loss = 29.551336, step = 418 (0.619 sec)\n",
"INFO:tensorflow:global_step/sec: 1.70913\n",
"INFO:tensorflow:train_accuracy = 0.49675745, train_loss = 25.094551 (0.585 sec)\n",
"INFO:tensorflow:loss = 25.094551, step = 419 (0.585 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64533\n",
"INFO:tensorflow:train_accuracy = 0.53642386, train_loss = 24.225819 (0.608 sec)\n",
"INFO:tensorflow:loss = 24.225819, step = 420 (0.608 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71297\n",
"INFO:tensorflow:train_accuracy = 0.5192308, train_loss = 25.992552 (0.584 sec)\n",
"INFO:tensorflow:loss = 25.992552, step = 421 (0.583 sec)\n",
"INFO:tensorflow:global_step/sec: 1.43691\n",
"INFO:tensorflow:train_accuracy = 0.50527704, train_loss = 25.722729 (0.696 sec)\n",
"INFO:tensorflow:loss = 25.722729, step = 422 (0.696 sec)\n",
"INFO:tensorflow:global_step/sec: 1.95404\n",
"INFO:tensorflow:train_accuracy = 0.5218543, train_loss = 24.503208 (0.512 sec)\n",
"INFO:tensorflow:loss = 24.503208, step = 423 (0.512 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65337\n",
"INFO:tensorflow:train_accuracy = 0.46031746, train_loss = 26.715132 (0.605 sec)\n",
"INFO:tensorflow:loss = 26.715132, step = 424 (0.605 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88886\n",
"INFO:tensorflow:train_accuracy = 0.5294118, train_loss = 24.458755 (0.529 sec)\n",
"INFO:tensorflow:loss = 24.458755, step = 425 (0.530 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77048\n",
"INFO:tensorflow:train_accuracy = 0.48181817, train_loss = 25.381657 (0.565 sec)\n",
"INFO:tensorflow:loss = 25.381657, step = 426 (0.565 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86251\n",
"INFO:tensorflow:train_accuracy = 0.5079576, train_loss = 24.94793 (0.537 sec)\n",
"INFO:tensorflow:loss = 24.94793, step = 427 (0.536 sec)\n",
"INFO:tensorflow:global_step/sec: 2.00393\n",
"INFO:tensorflow:train_accuracy = 0.5252918, train_loss = 25.30194 (0.499 sec)\n",
"INFO:tensorflow:loss = 25.30194, step = 428 (0.500 sec)\n",
"INFO:tensorflow:global_step/sec: 1.81349\n",
"INFO:tensorflow:train_accuracy = 0.5070785, train_loss = 25.651138 (0.551 sec)\n",
"INFO:tensorflow:loss = 25.651138, step = 429 (0.551 sec)\n",
"INFO:tensorflow:global_step/sec: 1.902\n",
"INFO:tensorflow:train_accuracy = 0.5164257, train_loss = 26.511347 (0.526 sec)\n",
"INFO:tensorflow:loss = 26.511347, step = 430 (0.526 sec)\n",
"INFO:tensorflow:global_step/sec: 2.04951\n",
"INFO:tensorflow:train_accuracy = 0.53856385, train_loss = 23.040365 (0.488 sec)\n",
"INFO:tensorflow:loss = 23.040365, step = 431 (0.488 sec)\n",
"INFO:tensorflow:global_step/sec: 2.03295\n",
"INFO:tensorflow:train_accuracy = 0.4993481, train_loss = 24.042202 (0.492 sec)\n",
"INFO:tensorflow:loss = 24.042202, step = 432 (0.492 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62904\n",
"INFO:tensorflow:train_accuracy = 0.49100256, train_loss = 28.942434 (0.614 sec)\n",
"INFO:tensorflow:loss = 28.942434, step = 433 (0.614 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73361\n",
"INFO:tensorflow:train_accuracy = 0.4897436, train_loss = 31.858027 (0.577 sec)\n",
"INFO:tensorflow:loss = 31.858027, step = 434 (0.577 sec)\n",
"INFO:tensorflow:global_step/sec: 1.78131\n",
"INFO:tensorflow:train_accuracy = 0.5097276, train_loss = 26.339859 (0.561 sec)\n",
"INFO:tensorflow:loss = 26.339859, step = 435 (0.561 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61718\n",
"INFO:tensorflow:train_accuracy = 0.5280313, train_loss = 27.339874 (0.619 sec)\n",
"INFO:tensorflow:loss = 27.339874, step = 436 (0.619 sec)\n",
"INFO:tensorflow:global_step/sec: 1.76798\n",
"INFO:tensorflow:train_accuracy = 0.5116883, train_loss = 24.34439 (0.566 sec)\n",
"INFO:tensorflow:loss = 24.34439, step = 437 (0.566 sec)\n",
"INFO:tensorflow:global_step/sec: 1.76721\n",
"INFO:tensorflow:train_accuracy = 0.5057915, train_loss = 28.266068 (0.566 sec)\n",
"INFO:tensorflow:loss = 28.266068, step = 438 (0.566 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86598\n",
"INFO:tensorflow:train_accuracy = 0.5378486, train_loss = 25.09143 (0.536 sec)\n",
"INFO:tensorflow:loss = 25.09143, step = 439 (0.535 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73336\n",
"INFO:tensorflow:train_accuracy = 0.4959893, train_loss = 26.120592 (0.577 sec)\n",
"INFO:tensorflow:loss = 26.120592, step = 440 (0.577 sec)\n",
"INFO:tensorflow:global_step/sec: 1.52587\n",
"INFO:tensorflow:train_accuracy = 0.48819876, train_loss = 32.755604 (0.655 sec)\n",
"INFO:tensorflow:loss = 32.755604, step = 441 (0.655 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64835\n",
"INFO:tensorflow:train_accuracy = 0.50448143, train_loss = 28.48369 (0.607 sec)\n",
"INFO:tensorflow:loss = 28.48369, step = 442 (0.606 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85738\n",
"INFO:tensorflow:train_accuracy = 0.47565544, train_loss = 33.393703 (0.539 sec)\n",
"INFO:tensorflow:loss = 33.393703, step = 443 (0.539 sec)\n",
"INFO:tensorflow:global_step/sec: 1.55609\n",
"INFO:tensorflow:train_accuracy = 0.50127226, train_loss = 29.266891 (0.643 sec)\n",
"INFO:tensorflow:loss = 29.266891, step = 444 (0.643 sec)\n",
"INFO:tensorflow:global_step/sec: 1.78122\n",
"INFO:tensorflow:train_accuracy = 0.5246338, train_loss = 22.249039 (0.561 sec)\n",
"INFO:tensorflow:loss = 22.249039, step = 445 (0.562 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77053\n",
"INFO:tensorflow:train_accuracy = 0.5246753, train_loss = 26.466055 (0.564 sec)\n",
"INFO:tensorflow:loss = 26.466055, step = 446 (0.563 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62519\n",
"INFO:tensorflow:train_accuracy = 0.48816937, train_loss = 28.376272 (0.616 sec)\n",
"INFO:tensorflow:loss = 28.376272, step = 447 (0.616 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86937\n",
"INFO:tensorflow:train_accuracy = 0.49234694, train_loss = 26.616055 (0.535 sec)\n",
"INFO:tensorflow:loss = 26.616055, step = 448 (0.534 sec)\n",
"INFO:tensorflow:global_step/sec: 1.70609\n",
"INFO:tensorflow:train_accuracy = 0.50879395, train_loss = 27.057213 (0.586 sec)\n",
"INFO:tensorflow:loss = 27.057213, step = 449 (0.585 sec)\n",
"INFO:tensorflow:global_step/sec: 1.5644\n",
"INFO:tensorflow:train_accuracy = 0.49815497, train_loss = 30.50175 (0.640 sec)\n",
"INFO:tensorflow:loss = 30.50175, step = 450 (0.640 sec)\n",
"INFO:tensorflow:global_step/sec: 2.01327\n",
"INFO:tensorflow:train_accuracy = 0.50448143, train_loss = 25.819311 (0.497 sec)\n",
"INFO:tensorflow:loss = 25.819311, step = 451 (0.497 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75621\n",
"INFO:tensorflow:train_accuracy = 0.5335968, train_loss = 24.041039 (0.569 sec)\n",
"INFO:tensorflow:loss = 24.041039, step = 452 (0.569 sec)\n",
"INFO:tensorflow:global_step/sec: 1.83828\n",
"INFO:tensorflow:train_accuracy = 0.50127226, train_loss = 27.252985 (0.544 sec)\n",
"INFO:tensorflow:loss = 27.252985, step = 453 (0.543 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71937\n",
"INFO:tensorflow:train_accuracy = 0.5101781, train_loss = 26.466608 (0.582 sec)\n",
"INFO:tensorflow:loss = 26.466608, step = 454 (0.582 sec)\n",
"INFO:tensorflow:global_step/sec: 1.78658\n",
"INFO:tensorflow:train_accuracy = 0.50448143, train_loss = 28.030807 (0.559 sec)\n",
"INFO:tensorflow:loss = 28.030807, step = 455 (0.559 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88043\n",
"INFO:tensorflow:train_accuracy = 0.5, train_loss = 26.692144 (0.532 sec)\n",
"INFO:tensorflow:loss = 26.692144, step = 456 (0.533 sec)\n",
"INFO:tensorflow:global_step/sec: 1.98523\n",
"INFO:tensorflow:train_accuracy = 0.5288591, train_loss = 23.427582 (0.504 sec)\n",
"INFO:tensorflow:loss = 23.427582, step = 457 (0.503 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88121\n",
"INFO:tensorflow:train_accuracy = 0.5302826, train_loss = 22.664307 (0.532 sec)\n",
"INFO:tensorflow:loss = 22.664307, step = 458 (0.532 sec)\n",
"INFO:tensorflow:global_step/sec: 5.36958\n",
"INFO:tensorflow:train_accuracy = 0.43457943, train_loss = 37.13166 (0.186 sec)\n",
"INFO:tensorflow:loss = 37.13166, step = 459 (0.186 sec)\n",
"INFO:tensorflow:global_step/sec: 1.6219\n",
"INFO:tensorflow:train_accuracy = 0.5032595, train_loss = 25.447773 (0.617 sec)\n",
"INFO:tensorflow:loss = 25.447773, step = 460 (0.617 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77737\n",
"INFO:tensorflow:train_accuracy = 0.49483204, train_loss = 26.06861 (0.563 sec)\n",
"INFO:tensorflow:loss = 26.06861, step = 461 (0.563 sec)\n",
"INFO:tensorflow:global_step/sec: 1.41735\n",
"INFO:tensorflow:train_accuracy = 0.43757576, train_loss = 36.671192 (0.706 sec)\n",
"INFO:tensorflow:loss = 36.671192, step = 462 (0.706 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71635\n",
"INFO:tensorflow:train_accuracy = 0.48451054, train_loss = 28.190994 (0.582 sec)\n",
"INFO:tensorflow:loss = 28.190994, step = 463 (0.582 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63651\n",
"INFO:tensorflow:train_accuracy = 0.4496124, train_loss = 31.61982 (0.611 sec)\n",
"INFO:tensorflow:loss = 31.61982, step = 464 (0.612 sec)\n",
"INFO:tensorflow:global_step/sec: 1.81482\n",
"INFO:tensorflow:train_accuracy = 0.515748, train_loss = 24.17671 (0.551 sec)\n",
"INFO:tensorflow:loss = 24.17671, step = 465 (0.551 sec)\n",
"INFO:tensorflow:global_step/sec: 1.68313\n",
"INFO:tensorflow:train_accuracy = 0.49673203, train_loss = 26.49918 (0.594 sec)\n",
"INFO:tensorflow:loss = 26.49918, step = 466 (0.595 sec)\n",
"INFO:tensorflow:global_step/sec: 1.68475\n",
"INFO:tensorflow:train_accuracy = 0.5246114, train_loss = 24.682957 (0.593 sec)\n",
"INFO:tensorflow:loss = 24.682957, step = 467 (0.593 sec)\n",
"INFO:tensorflow:global_step/sec: 1.6388\n",
"INFO:tensorflow:train_accuracy = 0.5232558, train_loss = 25.282639 (0.610 sec)\n",
"INFO:tensorflow:loss = 25.282639, step = 468 (0.610 sec)\n",
"INFO:tensorflow:global_step/sec: 1.45759\n",
"INFO:tensorflow:train_accuracy = 0.5060241, train_loss = 29.142511 (0.686 sec)\n",
"INFO:tensorflow:loss = 29.142511, step = 469 (0.685 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89418\n",
"INFO:tensorflow:train_accuracy = 0.486166, train_loss = 28.658607 (0.528 sec)\n",
"INFO:tensorflow:loss = 28.658607, step = 470 (0.528 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66238\n",
"INFO:tensorflow:train_accuracy = 0.5, train_loss = 26.426046 (0.602 sec)\n",
"INFO:tensorflow:loss = 26.426046, step = 471 (0.601 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91186\n",
"INFO:tensorflow:train_accuracy = 0.5164257, train_loss = 24.805635 (0.523 sec)\n",
"INFO:tensorflow:loss = 24.805635, step = 472 (0.523 sec)\n",
"INFO:tensorflow:global_step/sec: 1.9073\n",
"INFO:tensorflow:train_accuracy = 0.54289544, train_loss = 23.236656 (0.524 sec)\n",
"INFO:tensorflow:loss = 23.236656, step = 473 (0.524 sec)\n",
"INFO:tensorflow:global_step/sec: 2.02254\n",
"INFO:tensorflow:train_accuracy = 0.48157895, train_loss = 30.020536 (0.494 sec)\n",
"INFO:tensorflow:loss = 30.020536, step = 474 (0.495 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80004\n",
"INFO:tensorflow:train_accuracy = 0.5019157, train_loss = 26.153278 (0.556 sec)\n",
"INFO:tensorflow:loss = 26.153278, step = 475 (0.556 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87896\n",
"INFO:tensorflow:train_accuracy = 0.47809762, train_loss = 27.536472 (0.532 sec)\n",
"INFO:tensorflow:loss = 27.536472, step = 476 (0.532 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7274\n",
"INFO:tensorflow:train_accuracy = 0.48552632, train_loss = 25.81482 (0.579 sec)\n",
"INFO:tensorflow:loss = 25.81482, step = 477 (0.579 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86807\n",
"INFO:tensorflow:train_accuracy = 0.5123859, train_loss = 26.277826 (0.535 sec)\n",
"INFO:tensorflow:loss = 26.277826, step = 478 (0.535 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72712\n",
"INFO:tensorflow:train_accuracy = 0.5414438, train_loss = 24.282944 (0.579 sec)\n",
"INFO:tensorflow:loss = 24.282944, step = 479 (0.579 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82345\n",
"INFO:tensorflow:train_accuracy = 0.4915694, train_loss = 29.869205 (0.548 sec)\n",
"INFO:tensorflow:loss = 29.869205, step = 480 (0.549 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72938\n",
"INFO:tensorflow:train_accuracy = 0.5219072, train_loss = 27.084335 (0.578 sec)\n",
"INFO:tensorflow:loss = 27.084335, step = 481 (0.577 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7836\n",
"INFO:tensorflow:train_accuracy = 0.5280313, train_loss = 24.388676 (0.561 sec)\n",
"INFO:tensorflow:loss = 24.388676, step = 482 (0.562 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62193\n",
"INFO:tensorflow:train_accuracy = 0.51842105, train_loss = 24.62692 (0.617 sec)\n",
"INFO:tensorflow:loss = 24.62692, step = 483 (0.617 sec)\n",
"INFO:tensorflow:global_step/sec: 2.02479\n",
"INFO:tensorflow:train_accuracy = 0.52393275, train_loss = 24.137348 (0.494 sec)\n",
"INFO:tensorflow:loss = 24.137348, step = 484 (0.493 sec)\n",
"INFO:tensorflow:global_step/sec: 1.81073\n",
"INFO:tensorflow:train_accuracy = 0.5205298, train_loss = 27.138054 (0.552 sec)\n",
"INFO:tensorflow:loss = 27.138054, step = 485 (0.553 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72397\n",
"INFO:tensorflow:train_accuracy = 0.49870467, train_loss = 30.553219 (0.580 sec)\n",
"INFO:tensorflow:loss = 30.553219, step = 486 (0.580 sec)\n",
"INFO:tensorflow:global_step/sec: 1.52842\n",
"INFO:tensorflow:train_accuracy = 0.5081556, train_loss = 32.648346 (0.654 sec)\n",
"INFO:tensorflow:loss = 32.648346, step = 487 (0.654 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64161\n",
"INFO:tensorflow:train_accuracy = 0.5025575, train_loss = 25.073616 (0.609 sec)\n",
"INFO:tensorflow:loss = 25.073616, step = 488 (0.609 sec)\n",
"INFO:tensorflow:global_step/sec: 1.95849\n",
"INFO:tensorflow:train_accuracy = 0.50879395, train_loss = 27.462948 (0.511 sec)\n",
"INFO:tensorflow:loss = 27.462948, step = 489 (0.511 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90125\n",
"INFO:tensorflow:train_accuracy = 0.50463575, train_loss = 24.223598 (0.526 sec)\n",
"INFO:tensorflow:loss = 24.223598, step = 490 (0.526 sec)\n",
"INFO:tensorflow:global_step/sec: 1.51998\n",
"INFO:tensorflow:train_accuracy = 0.47875, train_loss = 32.90599 (0.658 sec)\n",
"INFO:tensorflow:loss = 32.90599, step = 491 (0.658 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7575\n",
"INFO:tensorflow:train_accuracy = 0.5150327, train_loss = 25.342407 (0.569 sec)\n",
"INFO:tensorflow:loss = 25.342407, step = 492 (0.569 sec)\n",
"INFO:tensorflow:global_step/sec: 1.78389\n",
"INFO:tensorflow:train_accuracy = 0.49620253, train_loss = 26.868744 (0.561 sec)\n",
"INFO:tensorflow:loss = 26.868744, step = 493 (0.561 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89596\n",
"INFO:tensorflow:train_accuracy = 0.5225806, train_loss = 25.593315 (0.527 sec)\n",
"INFO:tensorflow:loss = 25.593315, step = 494 (0.527 sec)\n",
"INFO:tensorflow:global_step/sec: 1.57419\n",
"INFO:tensorflow:train_accuracy = 0.4832714, train_loss = 28.135357 (0.635 sec)\n",
"INFO:tensorflow:loss = 28.135357, step = 495 (0.636 sec)\n",
"INFO:tensorflow:global_step/sec: 1.60226\n",
"INFO:tensorflow:train_accuracy = 0.49872774, train_loss = 27.105755 (0.624 sec)\n",
"INFO:tensorflow:loss = 27.105755, step = 496 (0.624 sec)\n",
"INFO:tensorflow:global_step/sec: 1.859\n",
"INFO:tensorflow:train_accuracy = 0.5019157, train_loss = 26.411703 (0.538 sec)\n",
"INFO:tensorflow:loss = 26.411703, step = 497 (0.538 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87277\n",
"INFO:tensorflow:train_accuracy = 0.49302915, train_loss = 27.21221 (0.534 sec)\n",
"INFO:tensorflow:loss = 27.21221, step = 498 (0.534 sec)\n",
"INFO:tensorflow:global_step/sec: 1.70316\n",
"INFO:tensorflow:train_accuracy = 0.50498754, train_loss = 28.730745 (0.587 sec)\n",
"INFO:tensorflow:loss = 28.730745, step = 499 (0.587 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80251\n",
"INFO:tensorflow:train_accuracy = 0.50199735, train_loss = 25.14109 (0.556 sec)\n",
"INFO:tensorflow:loss = 25.14109, step = 500 (0.556 sec)\n",
"INFO:tensorflow:global_step/sec: 1.947\n",
"INFO:tensorflow:train_accuracy = 0.5335968, train_loss = 23.915894 (0.513 sec)\n",
"INFO:tensorflow:loss = 23.915894, step = 501 (0.513 sec)\n",
"INFO:tensorflow:global_step/sec: 1.70287\n",
"INFO:tensorflow:train_accuracy = 0.47863248, train_loss = 30.427176 (0.587 sec)\n",
"INFO:tensorflow:loss = 30.427176, step = 502 (0.587 sec)\n",
"INFO:tensorflow:global_step/sec: 1.83956\n",
"INFO:tensorflow:train_accuracy = 0.5, train_loss = 24.280727 (0.543 sec)\n",
"INFO:tensorflow:loss = 24.280727, step = 503 (0.543 sec)\n",
"INFO:tensorflow:global_step/sec: 1.95268\n",
"INFO:tensorflow:train_accuracy = 0.54362416, train_loss = 20.97274 (0.513 sec)\n",
"INFO:tensorflow:loss = 20.97274, step = 504 (0.513 sec)\n",
"INFO:tensorflow:global_step/sec: 5.74167\n",
"INFO:tensorflow:train_accuracy = 0.5, train_loss = 24.207724 (0.174 sec)\n",
"INFO:tensorflow:loss = 24.207724, step = 505 (0.174 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63198\n",
"INFO:tensorflow:train_accuracy = 0.5246114, train_loss = 24.452055 (0.613 sec)\n",
"INFO:tensorflow:loss = 24.452055, step = 506 (0.613 sec)\n",
"INFO:tensorflow:global_step/sec: 1.68305\n",
"INFO:tensorflow:train_accuracy = 0.51007557, train_loss = 26.360039 (0.594 sec)\n",
"INFO:tensorflow:loss = 26.360039, step = 507 (0.594 sec)\n",
"INFO:tensorflow:global_step/sec: 1.40841\n",
"INFO:tensorflow:train_accuracy = 0.50498754, train_loss = 28.338783 (0.710 sec)\n",
"INFO:tensorflow:loss = 28.338783, step = 508 (0.710 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69005\n",
"INFO:tensorflow:train_accuracy = 0.48759305, train_loss = 30.701708 (0.591 sec)\n",
"INFO:tensorflow:loss = 30.701708, step = 509 (0.591 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90612\n",
"INFO:tensorflow:train_accuracy = 0.5378486, train_loss = 22.173151 (0.525 sec)\n",
"INFO:tensorflow:loss = 22.173151, step = 510 (0.526 sec)\n",
"INFO:tensorflow:global_step/sec: 1.6367\n",
"INFO:tensorflow:train_accuracy = 0.52056557, train_loss = 26.06724 (0.610 sec)\n",
"INFO:tensorflow:loss = 26.06724, step = 511 (0.610 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74741\n",
"INFO:tensorflow:train_accuracy = 0.52597404, train_loss = 24.108685 (0.573 sec)\n",
"INFO:tensorflow:loss = 24.108685, step = 512 (0.573 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66434\n",
"INFO:tensorflow:train_accuracy = 0.5130208, train_loss = 23.10094 (0.601 sec)\n",
"INFO:tensorflow:loss = 23.10094, step = 513 (0.600 sec)\n",
"INFO:tensorflow:global_step/sec: 1.67069\n",
"INFO:tensorflow:train_accuracy = 0.53149605, train_loss = 24.700932 (0.599 sec)\n",
"INFO:tensorflow:loss = 24.700932, step = 514 (0.599 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86435\n",
"INFO:tensorflow:train_accuracy = 0.5205298, train_loss = 24.507511 (0.536 sec)\n",
"INFO:tensorflow:loss = 24.507511, step = 515 (0.535 sec)\n",
"INFO:tensorflow:global_step/sec: 1.45555\n",
"INFO:tensorflow:train_accuracy = 0.515748, train_loss = 27.576134 (0.688 sec)\n",
"INFO:tensorflow:loss = 27.576134, step = 516 (0.688 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7268\n",
"INFO:tensorflow:train_accuracy = 0.53856385, train_loss = 23.956718 (0.579 sec)\n",
"INFO:tensorflow:loss = 23.956718, step = 517 (0.579 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91028\n",
"INFO:tensorflow:train_accuracy = 0.4948052, train_loss = 23.739185 (0.524 sec)\n",
"INFO:tensorflow:loss = 23.739185, step = 518 (0.524 sec)\n",
"INFO:tensorflow:global_step/sec: 1.67683\n",
"INFO:tensorflow:train_accuracy = 0.53149605, train_loss = 23.862709 (0.596 sec)\n",
"INFO:tensorflow:loss = 23.862709, step = 519 (0.596 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7566\n",
"INFO:tensorflow:train_accuracy = 0.4903226, train_loss = 26.589184 (0.569 sec)\n",
"INFO:tensorflow:loss = 26.589184, step = 520 (0.569 sec)\n",
"INFO:tensorflow:global_step/sec: 1.81551\n",
"INFO:tensorflow:train_accuracy = 0.49414825, train_loss = 25.38586 (0.551 sec)\n",
"INFO:tensorflow:loss = 25.38586, step = 521 (0.551 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80563\n",
"INFO:tensorflow:train_accuracy = 0.52734375, train_loss = 23.914928 (0.554 sec)\n",
"INFO:tensorflow:loss = 23.914928, step = 522 (0.554 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88999\n",
"INFO:tensorflow:train_accuracy = 0.50645995, train_loss = 24.228565 (0.529 sec)\n",
"INFO:tensorflow:loss = 24.228565, step = 523 (0.528 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89367\n",
"INFO:tensorflow:train_accuracy = 0.52047557, train_loss = 22.953373 (0.529 sec)\n",
"INFO:tensorflow:loss = 22.953373, step = 524 (0.529 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73704\n",
"INFO:tensorflow:train_accuracy = 0.5026455, train_loss = 27.30244 (0.575 sec)\n",
"INFO:tensorflow:loss = 27.30244, step = 525 (0.575 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61498\n",
"INFO:tensorflow:train_accuracy = 0.49125, train_loss = 28.955524 (0.619 sec)\n",
"INFO:tensorflow:loss = 28.955524, step = 526 (0.620 sec)\n",
"INFO:tensorflow:global_step/sec: 2.02807\n",
"INFO:tensorflow:train_accuracy = 0.50526315, train_loss = 25.211668 (0.493 sec)\n",
"INFO:tensorflow:loss = 25.211668, step = 527 (0.493 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77414\n",
"INFO:tensorflow:train_accuracy = 0.5246996, train_loss = 25.14355 (0.564 sec)\n",
"INFO:tensorflow:loss = 25.14355, step = 528 (0.564 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87007\n",
"INFO:tensorflow:train_accuracy = 0.50645995, train_loss = 25.182936 (0.534 sec)\n",
"INFO:tensorflow:loss = 25.182936, step = 529 (0.534 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85481\n",
"INFO:tensorflow:train_accuracy = 0.46946564, train_loss = 25.69622 (0.540 sec)\n",
"INFO:tensorflow:loss = 25.69622, step = 530 (0.540 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87636\n",
"INFO:tensorflow:train_accuracy = 0.5274431, train_loss = 23.32078 (0.532 sec)\n",
"INFO:tensorflow:loss = 23.32078, step = 531 (0.532 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82797\n",
"INFO:tensorflow:train_accuracy = 0.5197889, train_loss = 24.137081 (0.548 sec)\n",
"INFO:tensorflow:loss = 24.137081, step = 532 (0.548 sec)\n",
"INFO:tensorflow:global_step/sec: 1.50378\n",
"INFO:tensorflow:train_accuracy = 0.47571605, train_loss = 33.015366 (0.665 sec)\n",
"INFO:tensorflow:loss = 33.015366, step = 533 (0.665 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65792\n",
"INFO:tensorflow:train_accuracy = 0.5192308, train_loss = 25.271645 (0.603 sec)\n",
"INFO:tensorflow:loss = 25.271645, step = 534 (0.603 sec)\n",
"INFO:tensorflow:global_step/sec: 1.92668\n",
"INFO:tensorflow:train_accuracy = 0.4937186, train_loss = 28.164783 (0.519 sec)\n",
"INFO:tensorflow:loss = 28.164783, step = 535 (0.519 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87737\n",
"INFO:tensorflow:train_accuracy = 0.5157068, train_loss = 24.919569 (0.533 sec)\n",
"INFO:tensorflow:loss = 24.919569, step = 536 (0.533 sec)\n",
"INFO:tensorflow:global_step/sec: 1.50846\n",
"INFO:tensorflow:train_accuracy = 0.50564617, train_loss = 29.76904 (0.663 sec)\n",
"INFO:tensorflow:loss = 29.76904, step = 537 (0.662 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75762\n",
"INFO:tensorflow:train_accuracy = 0.5104167, train_loss = 23.571053 (0.569 sec)\n",
"INFO:tensorflow:loss = 23.571053, step = 538 (0.569 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71814\n",
"INFO:tensorflow:train_accuracy = 0.49557522, train_loss = 27.275177 (0.582 sec)\n",
"INFO:tensorflow:loss = 27.275177, step = 539 (0.582 sec)\n",
"INFO:tensorflow:global_step/sec: 2.03003\n",
"INFO:tensorflow:train_accuracy = 0.49935317, train_loss = 24.97601 (0.493 sec)\n",
"INFO:tensorflow:loss = 24.97601, step = 540 (0.493 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72403\n",
"INFO:tensorflow:train_accuracy = 0.48343557, train_loss = 28.92735 (0.580 sec)\n",
"INFO:tensorflow:loss = 28.92735, step = 541 (0.580 sec)\n",
"INFO:tensorflow:global_step/sec: 1.54553\n",
"INFO:tensorflow:train_accuracy = 0.4810606, train_loss = 28.810387 (0.647 sec)\n",
"INFO:tensorflow:loss = 28.810387, step = 542 (0.647 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8828\n",
"INFO:tensorflow:train_accuracy = 0.50449294, train_loss = 24.8396 (0.531 sec)\n",
"INFO:tensorflow:loss = 24.8396, step = 543 (0.531 sec)\n",
"INFO:tensorflow:global_step/sec: 1.92353\n",
"INFO:tensorflow:train_accuracy = 0.4986945, train_loss = 25.785069 (0.520 sec)\n",
"INFO:tensorflow:loss = 25.785069, step = 544 (0.520 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89081\n",
"INFO:tensorflow:train_accuracy = 0.5110533, train_loss = 24.19566 (0.529 sec)\n",
"INFO:tensorflow:loss = 24.19566, step = 545 (0.529 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71433\n",
"INFO:tensorflow:train_accuracy = 0.4775561, train_loss = 27.766176 (0.583 sec)\n",
"INFO:tensorflow:loss = 27.766176, step = 546 (0.583 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65055\n",
"INFO:tensorflow:train_accuracy = 0.5301047, train_loss = 24.628792 (0.606 sec)\n",
"INFO:tensorflow:loss = 24.628792, step = 547 (0.606 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87434\n",
"INFO:tensorflow:train_accuracy = 0.49753696, train_loss = 28.420532 (0.533 sec)\n",
"INFO:tensorflow:loss = 28.420532, step = 548 (0.533 sec)\n",
"INFO:tensorflow:global_step/sec: 1.83459\n",
"INFO:tensorflow:train_accuracy = 0.54072094, train_loss = 22.902195 (0.545 sec)\n",
"INFO:tensorflow:loss = 22.902195, step = 549 (0.545 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87714\n",
"INFO:tensorflow:train_accuracy = 0.50661373, train_loss = 23.737068 (0.532 sec)\n",
"INFO:tensorflow:loss = 23.737068, step = 550 (0.532 sec)\n",
"INFO:tensorflow:global_step/sec: 6.38688\n",
"INFO:tensorflow:train_accuracy = 0.5247525, train_loss = 19.745733 (0.156 sec)\n",
"INFO:tensorflow:loss = 19.745733, step = 551 (0.156 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71347\n",
"INFO:tensorflow:train_accuracy = 0.5070785, train_loss = 24.703426 (0.584 sec)\n",
"INFO:tensorflow:loss = 24.703426, step = 552 (0.585 sec)\n",
"INFO:tensorflow:global_step/sec: 1.84823\n",
"INFO:tensorflow:train_accuracy = 0.5246114, train_loss = 24.271797 (0.541 sec)\n",
"INFO:tensorflow:loss = 24.271797, step = 553 (0.540 sec)\n",
"INFO:tensorflow:global_step/sec: 1.41541\n",
"INFO:tensorflow:train_accuracy = 0.50980395, train_loss = 31.23441 (0.707 sec)\n",
"INFO:tensorflow:loss = 31.23441, step = 554 (0.707 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62415\n",
"INFO:tensorflow:train_accuracy = 0.481761, train_loss = 28.448448 (0.616 sec)\n",
"INFO:tensorflow:loss = 28.448448, step = 555 (0.616 sec)\n",
"INFO:tensorflow:global_step/sec: 1.60963\n",
"INFO:tensorflow:train_accuracy = 0.5280313, train_loss = 24.592247 (0.622 sec)\n",
"INFO:tensorflow:loss = 24.592247, step = 556 (0.622 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61495\n",
"INFO:tensorflow:train_accuracy = 0.49873418, train_loss = 26.339787 (0.619 sec)\n",
"INFO:tensorflow:loss = 26.339787, step = 557 (0.618 sec)\n",
"INFO:tensorflow:global_step/sec: 1.67846\n",
"INFO:tensorflow:train_accuracy = 0.5105541, train_loss = 22.927666 (0.596 sec)\n",
"INFO:tensorflow:loss = 22.927666, step = 558 (0.596 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85444\n",
"INFO:tensorflow:train_accuracy = 0.53149605, train_loss = 23.999989 (0.539 sec)\n",
"INFO:tensorflow:loss = 23.999989, step = 559 (0.540 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63792\n",
"INFO:tensorflow:train_accuracy = 0.51232165, train_loss = 23.898308 (0.611 sec)\n",
"INFO:tensorflow:loss = 23.898308, step = 560 (0.610 sec)\n",
"INFO:tensorflow:global_step/sec: 1.96157\n",
"INFO:tensorflow:train_accuracy = 0.5378486, train_loss = 21.404076 (0.510 sec)\n",
"INFO:tensorflow:loss = 21.404076, step = 561 (0.511 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74183\n",
"INFO:tensorflow:train_accuracy = 0.48751643, train_loss = 24.79837 (0.574 sec)\n",
"INFO:tensorflow:loss = 24.79837, step = 562 (0.573 sec)\n",
"INFO:tensorflow:global_step/sec: 1.60863\n",
"INFO:tensorflow:train_accuracy = 0.5371353, train_loss = 22.873795 (0.621 sec)\n",
"INFO:tensorflow:loss = 22.873795, step = 563 (0.621 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65984\n",
"INFO:tensorflow:train_accuracy = 0.51910406, train_loss = 24.682892 (0.603 sec)\n",
"INFO:tensorflow:loss = 24.682892, step = 564 (0.603 sec)\n",
"INFO:tensorflow:global_step/sec: 1.43329\n",
"INFO:tensorflow:train_accuracy = 0.4916345, train_loss = 27.669865 (0.698 sec)\n",
"INFO:tensorflow:loss = 27.669865, step = 565 (0.698 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82441\n",
"INFO:tensorflow:train_accuracy = 0.5308804, train_loss = 23.439981 (0.548 sec)\n",
"INFO:tensorflow:loss = 23.439981, step = 566 (0.548 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82583\n",
"INFO:tensorflow:train_accuracy = 0.507732, train_loss = 25.777939 (0.548 sec)\n",
"INFO:tensorflow:loss = 25.777939, step = 567 (0.548 sec)\n",
"INFO:tensorflow:global_step/sec: 1.884\n",
"INFO:tensorflow:train_accuracy = 0.5143603, train_loss = 23.094608 (0.531 sec)\n",
"INFO:tensorflow:loss = 23.094608, step = 568 (0.531 sec)\n",
"INFO:tensorflow:global_step/sec: 2.01912\n",
"INFO:tensorflow:train_accuracy = 0.50970244, train_loss = 25.762264 (0.495 sec)\n",
"INFO:tensorflow:loss = 25.762264, step = 569 (0.495 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74716\n",
"INFO:tensorflow:train_accuracy = 0.5288591, train_loss = 22.200676 (0.572 sec)\n",
"INFO:tensorflow:loss = 22.200676, step = 570 (0.572 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62048\n",
"INFO:tensorflow:train_accuracy = 0.513308, train_loss = 27.233334 (0.617 sec)\n",
"INFO:tensorflow:loss = 27.233334, step = 571 (0.617 sec)\n",
"INFO:tensorflow:global_step/sec: 1.5939\n",
"INFO:tensorflow:train_accuracy = 0.5113636, train_loss = 27.040249 (0.627 sec)\n",
"INFO:tensorflow:loss = 27.040249, step = 572 (0.627 sec)\n",
"INFO:tensorflow:global_step/sec: 1.81368\n",
"INFO:tensorflow:train_accuracy = 0.5246338, train_loss = 21.997625 (0.551 sec)\n",
"INFO:tensorflow:loss = 21.997625, step = 573 (0.551 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75216\n",
"INFO:tensorflow:train_accuracy = 0.49102563, train_loss = 25.710396 (0.571 sec)\n",
"INFO:tensorflow:loss = 25.710396, step = 574 (0.570 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88343\n",
"INFO:tensorflow:train_accuracy = 0.50384617, train_loss = 25.20202 (0.531 sec)\n",
"INFO:tensorflow:loss = 25.20202, step = 575 (0.531 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85161\n",
"INFO:tensorflow:train_accuracy = 0.54, train_loss = 22.055523 (0.540 sec)\n",
"INFO:tensorflow:loss = 22.055523, step = 576 (0.540 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82037\n",
"INFO:tensorflow:train_accuracy = 0.5198973, train_loss = 25.291191 (0.550 sec)\n",
"INFO:tensorflow:loss = 25.291191, step = 577 (0.550 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72077\n",
"INFO:tensorflow:train_accuracy = 0.5343915, train_loss = 22.593357 (0.581 sec)\n",
"INFO:tensorflow:loss = 22.593357, step = 578 (0.582 sec)\n",
"INFO:tensorflow:Done calling model_fn.\n",
"INFO:tensorflow:Starting evaluation at 2020-10-25T19:00:53Z\n",
"INFO:tensorflow:Graph was finalized.\n",
"INFO:tensorflow:Restoring parameters from asr-quartznet/model.ckpt-600\n",
"INFO:tensorflow:Running local_init_op.\n",
"INFO:tensorflow:Done running local_init_op.\n",
"INFO:tensorflow:Evaluation [1/10]\n",
"INFO:tensorflow:Evaluation [2/10]\n",
"INFO:tensorflow:Evaluation [3/10]\n",
"INFO:tensorflow:Evaluation [4/10]\n",
"INFO:tensorflow:Evaluation [5/10]\n",
"INFO:tensorflow:Evaluation [6/10]\n",
"INFO:tensorflow:Evaluation [7/10]\n",
"INFO:tensorflow:Evaluation [8/10]\n",
"INFO:tensorflow:Evaluation [9/10]\n",
"INFO:tensorflow:Evaluation [10/10]\n",
"INFO:tensorflow:Finished evaluation at 2020-10-25-19:01:01\n",
"INFO:tensorflow:Saving dict for global step 600: WER = 1.0, accuracy = 0.5245399, global_step = 600, loss = 23.889023\n",
"INFO:tensorflow:Saving 'checkpoint_path' summary for global step 600: asr-quartznet/model.ckpt-600\n",
"INFO:tensorflow:global_step/sec: 0.0567543\n",
"INFO:tensorflow:train_accuracy = 0.5, train_loss = 24.538128 (17.620 sec)\n",
"INFO:tensorflow:loss = 24.538128, step = 599 (17.619 sec)\n",
"INFO:tensorflow:global_step/sec: 1.39304\n",
"INFO:tensorflow:train_accuracy = 0.4894932, train_loss = 28.25082 (0.719 sec)\n",
"INFO:tensorflow:loss = 28.25082, step = 600 (0.720 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62705\n",
"INFO:tensorflow:train_accuracy = 0.46039605, train_loss = 33.89925 (0.614 sec)\n",
"INFO:tensorflow:loss = 33.89925, step = 601 (0.613 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91566\n",
"INFO:tensorflow:train_accuracy = 0.547651, train_loss = 20.188618 (0.522 sec)\n",
"INFO:tensorflow:loss = 20.188618, step = 602 (0.522 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64426\n",
"INFO:tensorflow:train_accuracy = 0.5019108, train_loss = 25.039825 (0.608 sec)\n",
"INFO:tensorflow:loss = 25.039825, step = 603 (0.608 sec)\n",
"INFO:tensorflow:global_step/sec: 1.70667\n",
"INFO:tensorflow:train_accuracy = 0.492126, train_loss = 24.490868 (0.586 sec)\n",
"INFO:tensorflow:loss = 24.490868, step = 604 (0.586 sec)\n",
"INFO:tensorflow:global_step/sec: 1.6959\n",
"INFO:tensorflow:train_accuracy = 0.5136541, train_loss = 23.922571 (0.590 sec)\n",
"INFO:tensorflow:loss = 23.922571, step = 605 (0.589 sec)\n",
"INFO:tensorflow:global_step/sec: 1.67246\n",
"INFO:tensorflow:train_accuracy = 0.521181, train_loss = 26.127655 (0.598 sec)\n",
"INFO:tensorflow:loss = 26.127655, step = 606 (0.598 sec)\n",
"INFO:tensorflow:global_step/sec: 1.45699\n",
"INFO:tensorflow:train_accuracy = 0.5274431, train_loss = 24.359705 (0.687 sec)\n",
"INFO:tensorflow:loss = 24.359705, step = 607 (0.687 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62679\n",
"INFO:tensorflow:train_accuracy = 0.5026316, train_loss = 24.894665 (0.614 sec)\n",
"INFO:tensorflow:loss = 24.894665, step = 608 (0.614 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82932\n",
"INFO:tensorflow:train_accuracy = 0.50970244, train_loss = 25.259876 (0.547 sec)\n",
"INFO:tensorflow:loss = 25.259876, step = 609 (0.547 sec)\n",
"INFO:tensorflow:global_step/sec: 1.92196\n",
"INFO:tensorflow:train_accuracy = 0.5157068, train_loss = 23.484772 (0.520 sec)\n",
"INFO:tensorflow:loss = 23.484772, step = 610 (0.520 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90942\n",
"INFO:tensorflow:train_accuracy = 0.48748353, train_loss = 24.932247 (0.524 sec)\n",
"INFO:tensorflow:loss = 24.932247, step = 611 (0.524 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7946\n",
"INFO:tensorflow:train_accuracy = 0.5495251, train_loss = 21.390022 (0.557 sec)\n",
"INFO:tensorflow:loss = 21.390022, step = 612 (0.558 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7353\n",
"INFO:tensorflow:train_accuracy = 0.4648241, train_loss = 29.24266 (0.576 sec)\n",
"INFO:tensorflow:loss = 29.24266, step = 613 (0.576 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75537\n",
"INFO:tensorflow:train_accuracy = 0.52, train_loss = 27.269152 (0.570 sec)\n",
"INFO:tensorflow:loss = 27.269152, step = 614 (0.570 sec)\n",
"INFO:tensorflow:global_step/sec: 1.96544\n",
"INFO:tensorflow:train_accuracy = 0.5280313, train_loss = 24.843098 (0.509 sec)\n",
"INFO:tensorflow:loss = 24.843098, step = 615 (0.510 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90738\n",
"INFO:tensorflow:train_accuracy = 0.54508746, train_loss = 21.04274 (0.523 sec)\n",
"INFO:tensorflow:loss = 21.04274, step = 616 (0.522 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86582\n",
"INFO:tensorflow:train_accuracy = 0.52122015, train_loss = 25.544395 (0.536 sec)\n",
"INFO:tensorflow:loss = 25.544395, step = 617 (0.537 sec)\n",
"INFO:tensorflow:global_step/sec: 1.78237\n",
"INFO:tensorflow:train_accuracy = 0.4925, train_loss = 27.814363 (0.561 sec)\n",
"INFO:tensorflow:loss = 27.814363, step = 618 (0.561 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63412\n",
"INFO:tensorflow:train_accuracy = 0.5126263, train_loss = 28.299248 (0.612 sec)\n",
"INFO:tensorflow:loss = 28.299248, step = 619 (0.612 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80255\n",
"INFO:tensorflow:train_accuracy = 0.5197889, train_loss = 23.886095 (0.555 sec)\n",
"INFO:tensorflow:loss = 23.886095, step = 620 (0.555 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85341\n",
"INFO:tensorflow:train_accuracy = 0.5554072, train_loss = 21.749067 (0.540 sec)\n",
"INFO:tensorflow:loss = 21.749067, step = 621 (0.540 sec)\n",
"INFO:tensorflow:global_step/sec: 1.83613\n",
"INFO:tensorflow:train_accuracy = 0.5444887, train_loss = 22.256107 (0.544 sec)\n",
"INFO:tensorflow:loss = 22.256107, step = 622 (0.544 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91986\n",
"INFO:tensorflow:train_accuracy = 0.5079576, train_loss = 23.821255 (0.521 sec)\n",
"INFO:tensorflow:loss = 23.821255, step = 623 (0.521 sec)\n",
"INFO:tensorflow:global_step/sec: 1.93007\n",
"INFO:tensorflow:train_accuracy = 0.5070785, train_loss = 25.013388 (0.518 sec)\n",
"INFO:tensorflow:loss = 25.013388, step = 624 (0.518 sec)\n",
"INFO:tensorflow:global_step/sec: 1.53666\n",
"INFO:tensorflow:train_accuracy = 0.53061223, train_loss = 31.095322 (0.651 sec)\n",
"INFO:tensorflow:loss = 31.095322, step = 625 (0.651 sec)\n",
"INFO:tensorflow:global_step/sec: 2.03578\n",
"INFO:tensorflow:train_accuracy = 0.5107459, train_loss = 28.170044 (0.491 sec)\n",
"INFO:tensorflow:loss = 28.170044, step = 626 (0.491 sec)\n",
"INFO:tensorflow:global_step/sec: 1.98788\n",
"INFO:tensorflow:train_accuracy = 0.47493735, train_loss = 27.222883 (0.503 sec)\n",
"INFO:tensorflow:loss = 27.222883, step = 627 (0.503 sec)\n",
"INFO:tensorflow:global_step/sec: 1.79061\n",
"INFO:tensorflow:train_accuracy = 0.54953766, train_loss = 23.379387 (0.558 sec)\n",
"INFO:tensorflow:loss = 23.379387, step = 628 (0.559 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65194\n",
"INFO:tensorflow:train_accuracy = 0.530719, train_loss = 24.50647 (0.605 sec)\n",
"INFO:tensorflow:loss = 24.50647, step = 629 (0.605 sec)\n",
"INFO:tensorflow:global_step/sec: 1.52374\n",
"INFO:tensorflow:train_accuracy = 0.5252525, train_loss = 26.697264 (0.656 sec)\n",
"INFO:tensorflow:loss = 26.697264, step = 630 (0.655 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74267\n",
"INFO:tensorflow:train_accuracy = 0.48816937, train_loss = 29.457796 (0.574 sec)\n",
"INFO:tensorflow:loss = 29.457796, step = 631 (0.574 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85345\n",
"INFO:tensorflow:train_accuracy = 0.50970244, train_loss = 23.6003 (0.540 sec)\n",
"INFO:tensorflow:loss = 23.6003, step = 632 (0.540 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63123\n",
"INFO:tensorflow:train_accuracy = 0.47985348, train_loss = 28.795628 (0.613 sec)\n",
"INFO:tensorflow:loss = 28.795628, step = 633 (0.613 sec)\n",
"INFO:tensorflow:global_step/sec: 1.76331\n",
"INFO:tensorflow:train_accuracy = 0.53128994, train_loss = 25.964132 (0.567 sec)\n",
"INFO:tensorflow:loss = 25.964132, step = 634 (0.568 sec)\n",
"INFO:tensorflow:global_step/sec: 1.57069\n",
"INFO:tensorflow:train_accuracy = 0.51005024, train_loss = 27.751024 (0.636 sec)\n",
"INFO:tensorflow:loss = 27.751024, step = 635 (0.636 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89621\n",
"INFO:tensorflow:train_accuracy = 0.54664916, train_loss = 22.855854 (0.528 sec)\n",
"INFO:tensorflow:loss = 22.855854, step = 636 (0.528 sec)\n",
"INFO:tensorflow:global_step/sec: 1.92604\n",
"INFO:tensorflow:train_accuracy = 0.53061223, train_loss = 24.026863 (0.519 sec)\n",
"INFO:tensorflow:loss = 24.026863, step = 637 (0.519 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72621\n",
"INFO:tensorflow:train_accuracy = 0.5212903, train_loss = 25.273582 (0.579 sec)\n",
"INFO:tensorflow:loss = 25.273582, step = 638 (0.579 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75864\n",
"INFO:tensorflow:train_accuracy = 0.5301047, train_loss = 24.42805 (0.568 sec)\n",
"INFO:tensorflow:loss = 24.42805, step = 639 (0.568 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89616\n",
"INFO:tensorflow:train_accuracy = 0.539559, train_loss = 21.777016 (0.527 sec)\n",
"INFO:tensorflow:loss = 21.777016, step = 640 (0.528 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75115\n",
"INFO:tensorflow:train_accuracy = 0.51461244, train_loss = 25.848991 (0.571 sec)\n",
"INFO:tensorflow:loss = 25.848991, step = 641 (0.572 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82213\n",
"INFO:tensorflow:train_accuracy = 0.5511596, train_loss = 21.990532 (0.549 sec)\n",
"INFO:tensorflow:loss = 21.990532, step = 642 (0.549 sec)\n",
"INFO:tensorflow:global_step/sec: 5.57553\n",
"INFO:tensorflow:train_accuracy = 0.46696034, train_loss = 27.079132 (0.179 sec)\n",
"INFO:tensorflow:loss = 27.079132, step = 643 (0.179 sec)\n",
"INFO:tensorflow:global_step/sec: 1.70353\n",
"INFO:tensorflow:train_accuracy = 0.55026454, train_loss = 21.587477 (0.587 sec)\n",
"INFO:tensorflow:loss = 21.587477, step = 644 (0.587 sec)\n",
"INFO:tensorflow:global_step/sec: 1.41817\n",
"INFO:tensorflow:train_accuracy = 0.5, train_loss = 29.037014 (0.705 sec)\n",
"INFO:tensorflow:loss = 29.037014, step = 645 (0.705 sec)\n",
"INFO:tensorflow:global_step/sec: 1.79234\n",
"INFO:tensorflow:train_accuracy = 0.51265824, train_loss = 26.57571 (0.559 sec)\n",
"INFO:tensorflow:loss = 26.57571, step = 646 (0.559 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80266\n",
"INFO:tensorflow:train_accuracy = 0.5285896, train_loss = 24.154705 (0.554 sec)\n",
"INFO:tensorflow:loss = 24.154705, step = 647 (0.554 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80088\n",
"INFO:tensorflow:train_accuracy = 0.5031447, train_loss = 27.356415 (0.555 sec)\n",
"INFO:tensorflow:loss = 27.356415, step = 648 (0.556 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74417\n",
"INFO:tensorflow:train_accuracy = 0.5363276, train_loss = 23.34724 (0.573 sec)\n",
"INFO:tensorflow:loss = 23.34724, step = 649 (0.573 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61007\n",
"INFO:tensorflow:train_accuracy = 0.50813514, train_loss = 26.448627 (0.621 sec)\n",
"INFO:tensorflow:loss = 26.448627, step = 650 (0.622 sec)\n",
"INFO:tensorflow:global_step/sec: 1.97115\n",
"INFO:tensorflow:train_accuracy = 0.5474934, train_loss = 24.601788 (0.507 sec)\n",
"INFO:tensorflow:loss = 24.601788, step = 651 (0.507 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64506\n",
"INFO:tensorflow:train_accuracy = 0.54350734, train_loss = 21.94363 (0.608 sec)\n",
"INFO:tensorflow:loss = 21.94363, step = 652 (0.608 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82319\n",
"INFO:tensorflow:train_accuracy = 0.502551, train_loss = 25.669079 (0.549 sec)\n",
"INFO:tensorflow:loss = 25.669079, step = 653 (0.549 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8748\n",
"INFO:tensorflow:train_accuracy = 0.52533334, train_loss = 21.238956 (0.533 sec)\n",
"INFO:tensorflow:loss = 21.238956, step = 654 (0.534 sec)\n",
"INFO:tensorflow:global_step/sec: 1.47395\n",
"INFO:tensorflow:train_accuracy = 0.5124836, train_loss = 22.669601 (0.678 sec)\n",
"INFO:tensorflow:loss = 22.669601, step = 655 (0.678 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7457\n",
"INFO:tensorflow:train_accuracy = 0.481715, train_loss = 29.335497 (0.573 sec)\n",
"INFO:tensorflow:loss = 29.335497, step = 656 (0.573 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90221\n",
"INFO:tensorflow:train_accuracy = 0.5197279, train_loss = 23.239876 (0.525 sec)\n",
"INFO:tensorflow:loss = 23.239876, step = 657 (0.525 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86511\n",
"INFO:tensorflow:train_accuracy = 0.5287206, train_loss = 23.994373 (0.536 sec)\n",
"INFO:tensorflow:loss = 23.994373, step = 658 (0.536 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65064\n",
"INFO:tensorflow:train_accuracy = 0.513308, train_loss = 25.994253 (0.606 sec)\n",
"INFO:tensorflow:loss = 25.994253, step = 659 (0.607 sec)\n",
"INFO:tensorflow:global_step/sec: 2.01108\n",
"INFO:tensorflow:train_accuracy = 0.51706034, train_loss = 25.102896 (0.497 sec)\n",
"INFO:tensorflow:loss = 25.102896, step = 660 (0.497 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86923\n",
"INFO:tensorflow:train_accuracy = 0.51778656, train_loss = 24.953617 (0.535 sec)\n",
"INFO:tensorflow:loss = 24.953617, step = 661 (0.535 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80051\n",
"INFO:tensorflow:train_accuracy = 0.53532606, train_loss = 21.749018 (0.556 sec)\n",
"INFO:tensorflow:loss = 21.749018, step = 662 (0.556 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71517\n",
"INFO:tensorflow:train_accuracy = 0.4816223, train_loss = 28.358248 (0.583 sec)\n",
"INFO:tensorflow:loss = 28.358248, step = 663 (0.583 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62204\n",
"INFO:tensorflow:train_accuracy = 0.49810845, train_loss = 32.553257 (0.616 sec)\n",
"INFO:tensorflow:loss = 32.553257, step = 664 (0.616 sec)\n",
"INFO:tensorflow:global_step/sec: 1.95081\n",
"INFO:tensorflow:train_accuracy = 0.5103627, train_loss = 22.817604 (0.513 sec)\n",
"INFO:tensorflow:loss = 22.817604, step = 665 (0.513 sec)\n",
"INFO:tensorflow:global_step/sec: 1.801\n",
"INFO:tensorflow:train_accuracy = 0.5137255, train_loss = 23.669085 (0.555 sec)\n",
"INFO:tensorflow:loss = 23.669085, step = 666 (0.555 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86579\n",
"INFO:tensorflow:train_accuracy = 0.53642386, train_loss = 22.65345 (0.536 sec)\n",
"INFO:tensorflow:loss = 22.65345, step = 667 (0.536 sec)\n",
"INFO:tensorflow:global_step/sec: 1.99141\n",
"INFO:tensorflow:train_accuracy = 0.54508746, train_loss = 21.66401 (0.502 sec)\n",
"INFO:tensorflow:loss = 21.66401, step = 668 (0.502 sec)\n",
"INFO:tensorflow:global_step/sec: 1.78025\n",
"INFO:tensorflow:train_accuracy = 0.4937656, train_loss = 28.83725 (0.562 sec)\n",
"INFO:tensorflow:loss = 28.83725, step = 669 (0.562 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71666\n",
"INFO:tensorflow:train_accuracy = 0.5419532, train_loss = 20.867577 (0.583 sec)\n",
"INFO:tensorflow:loss = 20.867577, step = 670 (0.583 sec)\n",
"INFO:tensorflow:global_step/sec: 1.70929\n",
"INFO:tensorflow:train_accuracy = 0.5139594, train_loss = 25.517475 (0.585 sec)\n",
"INFO:tensorflow:loss = 25.517475, step = 671 (0.585 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64546\n",
"INFO:tensorflow:train_accuracy = 0.5138889, train_loss = 25.68267 (0.607 sec)\n",
"INFO:tensorflow:loss = 25.68267, step = 672 (0.607 sec)\n",
"INFO:tensorflow:global_step/sec: 1.54513\n",
"INFO:tensorflow:train_accuracy = 0.52791876, train_loss = 29.278015 (0.647 sec)\n",
"INFO:tensorflow:loss = 29.278015, step = 673 (0.647 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86906\n",
"INFO:tensorflow:train_accuracy = 0.5025575, train_loss = 26.226906 (0.535 sec)\n",
"INFO:tensorflow:loss = 26.226906, step = 674 (0.535 sec)\n",
"INFO:tensorflow:global_step/sec: 1.50255\n",
"INFO:tensorflow:train_accuracy = 0.52911395, train_loss = 28.265034 (0.666 sec)\n",
"INFO:tensorflow:loss = 28.265034, step = 675 (0.666 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75103\n",
"INFO:tensorflow:train_accuracy = 0.53289473, train_loss = 24.23627 (0.571 sec)\n",
"INFO:tensorflow:loss = 24.23627, step = 676 (0.571 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75721\n",
"INFO:tensorflow:train_accuracy = 0.49127182, train_loss = 26.573174 (0.570 sec)\n",
"INFO:tensorflow:loss = 26.573174, step = 677 (0.569 sec)\n",
"INFO:tensorflow:global_step/sec: 1.76053\n",
"INFO:tensorflow:train_accuracy = 0.5239948, train_loss = 24.539417 (0.567 sec)\n",
"INFO:tensorflow:loss = 24.539417, step = 678 (0.567 sec)\n",
"INFO:tensorflow:global_step/sec: 1.56345\n",
"INFO:tensorflow:train_accuracy = 0.50625, train_loss = 27.704393 (0.640 sec)\n",
"INFO:tensorflow:loss = 27.704393, step = 679 (0.639 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77406\n",
"INFO:tensorflow:train_accuracy = 0.49061325, train_loss = 28.207188 (0.564 sec)\n",
"INFO:tensorflow:loss = 28.207188, step = 680 (0.564 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73982\n",
"INFO:tensorflow:train_accuracy = 0.49129352, train_loss = 28.763418 (0.575 sec)\n",
"INFO:tensorflow:loss = 28.763418, step = 681 (0.575 sec)\n",
"INFO:tensorflow:global_step/sec: 1.93609\n",
"INFO:tensorflow:train_accuracy = 0.5084306, train_loss = 22.608562 (0.516 sec)\n",
"INFO:tensorflow:loss = 22.608562, step = 682 (0.517 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62242\n",
"INFO:tensorflow:train_accuracy = 0.5252918, train_loss = 25.896124 (0.617 sec)\n",
"INFO:tensorflow:loss = 25.896124, step = 683 (0.616 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72795\n",
"INFO:tensorflow:train_accuracy = 0.52948886, train_loss = 25.096586 (0.578 sec)\n",
"INFO:tensorflow:loss = 25.096586, step = 684 (0.578 sec)\n",
"INFO:tensorflow:global_step/sec: 1.79879\n",
"INFO:tensorflow:train_accuracy = 0.50709677, train_loss = 26.87746 (0.556 sec)\n",
"INFO:tensorflow:loss = 26.87746, step = 685 (0.556 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82224\n",
"INFO:tensorflow:train_accuracy = 0.49815497, train_loss = 28.314005 (0.548 sec)\n",
"INFO:tensorflow:loss = 28.314005, step = 686 (0.548 sec)\n",
"INFO:tensorflow:global_step/sec: 2.05406\n",
"INFO:tensorflow:train_accuracy = 0.5586207, train_loss = 19.044138 (0.487 sec)\n",
"INFO:tensorflow:loss = 19.044138, step = 687 (0.488 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87858\n",
"INFO:tensorflow:train_accuracy = 0.5294118, train_loss = 24.86093 (0.533 sec)\n",
"INFO:tensorflow:loss = 24.86093, step = 688 (0.533 sec)\n",
"INFO:tensorflow:global_step/sec: 4.60575\n",
"INFO:tensorflow:train_accuracy = 0.35779816, train_loss = 33.159588 (0.216 sec)\n",
"INFO:tensorflow:loss = 33.159588, step = 689 (0.216 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66738\n",
"INFO:tensorflow:train_accuracy = 0.5088161, train_loss = 27.773605 (0.600 sec)\n",
"INFO:tensorflow:loss = 27.773605, step = 690 (0.600 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64441\n",
"INFO:tensorflow:train_accuracy = 0.5110821, train_loss = 22.494053 (0.608 sec)\n",
"INFO:tensorflow:loss = 22.494053, step = 691 (0.608 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66072\n",
"INFO:tensorflow:train_accuracy = 0.49, train_loss = 28.171738 (0.602 sec)\n",
"INFO:tensorflow:loss = 28.171738, step = 692 (0.602 sec)\n",
"INFO:tensorflow:global_step/sec: 1.40997\n",
"INFO:tensorflow:train_accuracy = 0.47619048, train_loss = 31.452408 (0.709 sec)\n",
"INFO:tensorflow:loss = 31.452408, step = 693 (0.709 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90556\n",
"INFO:tensorflow:train_accuracy = 0.5465587, train_loss = 22.813534 (0.524 sec)\n",
"INFO:tensorflow:loss = 22.813534, step = 694 (0.524 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62909\n",
"INFO:tensorflow:train_accuracy = 0.49746835, train_loss = 25.357101 (0.614 sec)\n",
"INFO:tensorflow:loss = 25.357101, step = 695 (0.614 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69458\n",
"INFO:tensorflow:train_accuracy = 0.5137255, train_loss = 24.349255 (0.590 sec)\n",
"INFO:tensorflow:loss = 24.349255, step = 696 (0.590 sec)\n",
"INFO:tensorflow:global_step/sec: 1.93714\n",
"INFO:tensorflow:train_accuracy = 0.5013089, train_loss = 23.587101 (0.516 sec)\n",
"INFO:tensorflow:loss = 23.587101, step = 697 (0.515 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63162\n",
"INFO:tensorflow:train_accuracy = 0.496134, train_loss = 25.485226 (0.613 sec)\n",
"INFO:tensorflow:loss = 25.485226, step = 698 (0.613 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90189\n",
"INFO:tensorflow:train_accuracy = 0.53642386, train_loss = 24.096952 (0.526 sec)\n",
"INFO:tensorflow:loss = 24.096952, step = 699 (0.526 sec)\n",
"INFO:tensorflow:global_step/sec: 1.4437\n",
"INFO:tensorflow:train_accuracy = 0.5197889, train_loss = 26.283838 (0.694 sec)\n",
"INFO:tensorflow:loss = 26.283838, step = 700 (0.694 sec)\n",
"INFO:tensorflow:global_step/sec: 1.6725\n",
"INFO:tensorflow:train_accuracy = 0.50651044, train_loss = 23.528786 (0.596 sec)\n",
"INFO:tensorflow:loss = 23.528786, step = 701 (0.596 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87864\n",
"INFO:tensorflow:train_accuracy = 0.53289473, train_loss = 23.671024 (0.533 sec)\n",
"INFO:tensorflow:loss = 23.671024, step = 702 (0.533 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91845\n",
"INFO:tensorflow:train_accuracy = 0.51428574, train_loss = 24.293999 (0.521 sec)\n",
"INFO:tensorflow:loss = 24.293999, step = 703 (0.521 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89061\n",
"INFO:tensorflow:train_accuracy = 0.52393275, train_loss = 23.984558 (0.529 sec)\n",
"INFO:tensorflow:loss = 23.984558, step = 704 (0.530 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77534\n",
"INFO:tensorflow:train_accuracy = 0.53289473, train_loss = 23.865236 (0.563 sec)\n",
"INFO:tensorflow:loss = 23.865236, step = 705 (0.562 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63614\n",
"INFO:tensorflow:train_accuracy = 0.5219072, train_loss = 24.729189 (0.612 sec)\n",
"INFO:tensorflow:loss = 24.729189, step = 706 (0.612 sec)\n",
"INFO:tensorflow:global_step/sec: 2.08827\n",
"INFO:tensorflow:train_accuracy = 0.53079945, train_loss = 23.368896 (0.478 sec)\n",
"INFO:tensorflow:loss = 23.368896, step = 707 (0.478 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80141\n",
"INFO:tensorflow:train_accuracy = 0.5124183, train_loss = 23.665169 (0.555 sec)\n",
"INFO:tensorflow:loss = 23.665169, step = 708 (0.555 sec)\n",
"INFO:tensorflow:global_step/sec: 1.78935\n",
"INFO:tensorflow:train_accuracy = 0.49037227, train_loss = 27.187862 (0.559 sec)\n",
"INFO:tensorflow:loss = 27.187862, step = 709 (0.558 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90663\n",
"INFO:tensorflow:train_accuracy = 0.53856385, train_loss = 21.67915 (0.524 sec)\n",
"INFO:tensorflow:loss = 21.67915, step = 710 (0.524 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8297\n",
"INFO:tensorflow:train_accuracy = 0.54072094, train_loss = 21.601942 (0.547 sec)\n",
"INFO:tensorflow:loss = 21.601942, step = 711 (0.548 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61144\n",
"INFO:tensorflow:train_accuracy = 0.50127226, train_loss = 26.373356 (0.620 sec)\n",
"INFO:tensorflow:loss = 26.373356, step = 712 (0.620 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75847\n",
"INFO:tensorflow:train_accuracy = 0.5192308, train_loss = 24.99762 (0.569 sec)\n",
"INFO:tensorflow:loss = 24.99762, step = 713 (0.569 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87453\n",
"INFO:tensorflow:train_accuracy = 0.53856385, train_loss = 21.04975 (0.533 sec)\n",
"INFO:tensorflow:loss = 21.04975, step = 714 (0.533 sec)\n",
"INFO:tensorflow:global_step/sec: 1.92477\n",
"INFO:tensorflow:train_accuracy = 0.49419355, train_loss = 26.727999 (0.520 sec)\n",
"INFO:tensorflow:loss = 26.727999, step = 715 (0.520 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61868\n",
"INFO:tensorflow:train_accuracy = 0.54350734, train_loss = 22.113716 (0.618 sec)\n",
"INFO:tensorflow:loss = 22.113716, step = 716 (0.617 sec)\n",
"INFO:tensorflow:global_step/sec: 1.52225\n",
"INFO:tensorflow:train_accuracy = 0.50449294, train_loss = 31.41482 (0.657 sec)\n",
"INFO:tensorflow:loss = 31.41482, step = 717 (0.658 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72288\n",
"INFO:tensorflow:train_accuracy = 0.49632353, train_loss = 28.431633 (0.580 sec)\n",
"INFO:tensorflow:loss = 28.431633, step = 718 (0.580 sec)\n",
"INFO:tensorflow:global_step/sec: 1.97344\n",
"INFO:tensorflow:train_accuracy = 0.50642675, train_loss = 24.341 (0.507 sec)\n",
"INFO:tensorflow:loss = 24.341, step = 719 (0.508 sec)\n",
"INFO:tensorflow:global_step/sec: 1.99192\n",
"INFO:tensorflow:train_accuracy = 0.51910406, train_loss = 24.349861 (0.502 sec)\n",
"INFO:tensorflow:loss = 24.349861, step = 720 (0.501 sec)\n",
"INFO:tensorflow:global_step/sec: 1.97095\n",
"INFO:tensorflow:train_accuracy = 0.5205128, train_loss = 24.87104 (0.508 sec)\n",
"INFO:tensorflow:loss = 24.87104, step = 721 (0.507 sec)\n",
"INFO:tensorflow:global_step/sec: 1.52826\n",
"INFO:tensorflow:train_accuracy = 0.5044586, train_loss = 31.80988 (0.655 sec)\n",
"INFO:tensorflow:loss = 31.80988, step = 722 (0.655 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66957\n",
"INFO:tensorflow:train_accuracy = 0.5249042, train_loss = 26.769995 (0.599 sec)\n",
"INFO:tensorflow:loss = 26.769995, step = 723 (0.598 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87404\n",
"INFO:tensorflow:train_accuracy = 0.510665, train_loss = 27.34491 (0.534 sec)\n",
"INFO:tensorflow:loss = 27.34491, step = 724 (0.534 sec)\n",
"INFO:tensorflow:global_step/sec: 1.56829\n",
"INFO:tensorflow:train_accuracy = 0.4925187, train_loss = 27.125954 (0.638 sec)\n",
"INFO:tensorflow:loss = 27.125954, step = 725 (0.638 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75126\n",
"INFO:tensorflow:train_accuracy = 0.52387094, train_loss = 24.700418 (0.571 sec)\n",
"INFO:tensorflow:loss = 24.700418, step = 726 (0.571 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72527\n",
"INFO:tensorflow:train_accuracy = 0.47994986, train_loss = 30.153385 (0.580 sec)\n",
"INFO:tensorflow:loss = 30.153385, step = 727 (0.580 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89566\n",
"INFO:tensorflow:train_accuracy = 0.52590674, train_loss = 22.201553 (0.527 sec)\n",
"INFO:tensorflow:loss = 22.201553, step = 728 (0.527 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80263\n",
"INFO:tensorflow:train_accuracy = 0.5185185, train_loss = 26.272478 (0.555 sec)\n",
"INFO:tensorflow:loss = 26.272478, step = 729 (0.555 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71024\n",
"INFO:tensorflow:train_accuracy = 0.509434, train_loss = 27.21032 (0.584 sec)\n",
"INFO:tensorflow:loss = 27.21032, step = 730 (0.584 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75704\n",
"INFO:tensorflow:train_accuracy = 0.50771207, train_loss = 27.490623 (0.569 sec)\n",
"INFO:tensorflow:loss = 27.490623, step = 731 (0.570 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91326\n",
"INFO:tensorflow:train_accuracy = 0.50127226, train_loss = 24.056564 (0.523 sec)\n",
"INFO:tensorflow:loss = 24.056564, step = 732 (0.523 sec)\n",
"INFO:tensorflow:global_step/sec: 1.9119\n",
"INFO:tensorflow:train_accuracy = 0.5421687, train_loss = 22.604435 (0.523 sec)\n",
"INFO:tensorflow:loss = 22.604435, step = 733 (0.524 sec)\n",
"INFO:tensorflow:global_step/sec: 1.93655\n",
"INFO:tensorflow:train_accuracy = 0.54362416, train_loss = 20.732424 (0.516 sec)\n",
"INFO:tensorflow:loss = 20.732424, step = 734 (0.516 sec)\n",
"INFO:tensorflow:global_step/sec: 5.81574\n",
"INFO:tensorflow:train_accuracy = 0.49302325, train_loss = 27.793705 (0.172 sec)\n",
"INFO:tensorflow:loss = 27.793705, step = 735 (0.172 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71835\n",
"INFO:tensorflow:train_accuracy = 0.5162162, train_loss = 20.561872 (0.582 sec)\n",
"INFO:tensorflow:loss = 20.561872, step = 736 (0.583 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62406\n",
"INFO:tensorflow:train_accuracy = 0.49571604, train_loss = 29.777832 (0.615 sec)\n",
"INFO:tensorflow:loss = 29.777832, step = 737 (0.614 sec)\n",
"INFO:tensorflow:global_step/sec: 1.78218\n",
"INFO:tensorflow:train_accuracy = 0.49873418, train_loss = 25.782772 (0.561 sec)\n",
"INFO:tensorflow:loss = 25.782772, step = 738 (0.562 sec)\n",
"INFO:tensorflow:global_step/sec: 1.42448\n",
"INFO:tensorflow:train_accuracy = 0.4437577, train_loss = 34.78264 (0.702 sec)\n",
"INFO:tensorflow:loss = 34.78264, step = 739 (0.703 sec)\n",
"INFO:tensorflow:global_step/sec: 1.6209\n",
"INFO:tensorflow:train_accuracy = 0.51658165, train_loss = 25.395088 (0.617 sec)\n",
"INFO:tensorflow:loss = 25.395088, step = 740 (0.616 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65859\n",
"INFO:tensorflow:train_accuracy = 0.5096774, train_loss = 24.051954 (0.603 sec)\n",
"INFO:tensorflow:loss = 24.051954, step = 741 (0.603 sec)\n",
"INFO:tensorflow:global_step/sec: 1.863\n",
"INFO:tensorflow:train_accuracy = 0.5143979, train_loss = 25.175915 (0.537 sec)\n",
"INFO:tensorflow:loss = 25.175915, step = 742 (0.538 sec)\n",
"INFO:tensorflow:global_step/sec: 1.95828\n",
"INFO:tensorflow:train_accuracy = 0.5458221, train_loss = 21.386982 (0.511 sec)\n",
"INFO:tensorflow:loss = 21.386982, step = 743 (0.511 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63891\n",
"INFO:tensorflow:train_accuracy = 0.49357328, train_loss = 24.95757 (0.609 sec)\n",
"INFO:tensorflow:loss = 24.95757, step = 744 (0.609 sec)\n",
"INFO:tensorflow:global_step/sec: 1.46324\n",
"INFO:tensorflow:train_accuracy = 0.4922879, train_loss = 25.872192 (0.683 sec)\n",
"INFO:tensorflow:loss = 25.872192, step = 745 (0.683 sec)\n",
"INFO:tensorflow:global_step/sec: 1.982\n",
"INFO:tensorflow:train_accuracy = 0.52956986, train_loss = 22.34168 (0.505 sec)\n",
"INFO:tensorflow:loss = 22.34168, step = 746 (0.505 sec)\n",
"INFO:tensorflow:global_step/sec: 1.97187\n",
"INFO:tensorflow:train_accuracy = 0.5197889, train_loss = 23.73894 (0.507 sec)\n",
"INFO:tensorflow:loss = 23.73894, step = 747 (0.507 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66135\n",
"INFO:tensorflow:train_accuracy = 0.5136541, train_loss = 26.24224 (0.602 sec)\n",
"INFO:tensorflow:loss = 26.24224, step = 748 (0.602 sec)\n",
"INFO:tensorflow:global_step/sec: 1.84107\n",
"INFO:tensorflow:train_accuracy = 0.5335968, train_loss = 22.55331 (0.543 sec)\n",
"INFO:tensorflow:loss = 22.55331, step = 749 (0.543 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90586\n",
"INFO:tensorflow:train_accuracy = 0.5163827, train_loss = 23.033218 (0.525 sec)\n",
"INFO:tensorflow:loss = 23.033218, step = 750 (0.525 sec)\n",
"INFO:tensorflow:global_step/sec: 2.1041\n",
"INFO:tensorflow:train_accuracy = 0.51861364, train_loss = 24.349098 (0.475 sec)\n",
"INFO:tensorflow:loss = 24.349098, step = 751 (0.474 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86916\n",
"INFO:tensorflow:train_accuracy = 0.5218509, train_loss = 25.71922 (0.536 sec)\n",
"INFO:tensorflow:loss = 25.71922, step = 752 (0.536 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85606\n",
"INFO:tensorflow:train_accuracy = 0.50879395, train_loss = 26.216736 (0.539 sec)\n",
"INFO:tensorflow:loss = 26.216736, step = 753 (0.538 sec)\n",
"INFO:tensorflow:global_step/sec: 2.02108\n",
"INFO:tensorflow:train_accuracy = 0.54435486, train_loss = 22.181927 (0.495 sec)\n",
"INFO:tensorflow:loss = 22.181927, step = 754 (0.495 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8067\n",
"INFO:tensorflow:train_accuracy = 0.5317139, train_loss = 23.759045 (0.553 sec)\n",
"INFO:tensorflow:loss = 23.759045, step = 755 (0.553 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62694\n",
"INFO:tensorflow:train_accuracy = 0.5128866, train_loss = 25.01212 (0.615 sec)\n",
"INFO:tensorflow:loss = 25.01212, step = 756 (0.615 sec)\n",
"INFO:tensorflow:global_step/sec: 1.79364\n",
"INFO:tensorflow:train_accuracy = 0.5081556, train_loss = 26.937231 (0.557 sec)\n",
"INFO:tensorflow:loss = 26.937231, step = 757 (0.557 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7635\n",
"INFO:tensorflow:train_accuracy = 0.5032938, train_loss = 23.799713 (0.567 sec)\n",
"INFO:tensorflow:loss = 23.799713, step = 758 (0.567 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8879\n",
"INFO:tensorflow:train_accuracy = 0.5252918, train_loss = 23.613052 (0.530 sec)\n",
"INFO:tensorflow:loss = 23.613052, step = 759 (0.530 sec)\n",
"INFO:tensorflow:global_step/sec: 1.98968\n",
"INFO:tensorflow:train_accuracy = 0.5065274, train_loss = 25.306925 (0.503 sec)\n",
"INFO:tensorflow:loss = 25.306925, step = 760 (0.502 sec)\n",
"INFO:tensorflow:global_step/sec: 1.79461\n",
"INFO:tensorflow:train_accuracy = 0.53642386, train_loss = 24.907063 (0.557 sec)\n",
"INFO:tensorflow:loss = 24.907063, step = 761 (0.558 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73151\n",
"INFO:tensorflow:train_accuracy = 0.5369393, train_loss = 22.249722 (0.577 sec)\n",
"INFO:tensorflow:loss = 22.249722, step = 762 (0.577 sec)\n",
"INFO:tensorflow:global_step/sec: 1.55127\n",
"INFO:tensorflow:train_accuracy = 0.5, train_loss = 29.601543 (0.644 sec)\n",
"INFO:tensorflow:loss = 29.601543, step = 763 (0.644 sec)\n",
"INFO:tensorflow:global_step/sec: 1.9681\n",
"INFO:tensorflow:train_accuracy = 0.5246114, train_loss = 22.571758 (0.508 sec)\n",
"INFO:tensorflow:loss = 22.571758, step = 764 (0.508 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91813\n",
"INFO:tensorflow:train_accuracy = 0.50061655, train_loss = 29.55262 (0.521 sec)\n",
"INFO:tensorflow:loss = 29.55262, step = 765 (0.521 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74364\n",
"INFO:tensorflow:train_accuracy = 0.5104167, train_loss = 24.878258 (0.573 sec)\n",
"INFO:tensorflow:loss = 24.878258, step = 766 (0.573 sec)\n",
"INFO:tensorflow:global_step/sec: 1.53375\n",
"INFO:tensorflow:train_accuracy = 0.53238344, train_loss = 32.50962 (0.652 sec)\n",
"INFO:tensorflow:loss = 32.50962, step = 767 (0.652 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61969\n",
"INFO:tensorflow:train_accuracy = 0.51772153, train_loss = 25.498552 (0.618 sec)\n",
"INFO:tensorflow:loss = 25.498552, step = 768 (0.618 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7343\n",
"INFO:tensorflow:train_accuracy = 0.525, train_loss = 26.295805 (0.576 sec)\n",
"INFO:tensorflow:loss = 26.295805, step = 769 (0.576 sec)\n",
"INFO:tensorflow:global_step/sec: 1.87135\n",
"INFO:tensorflow:train_accuracy = 0.54509807, train_loss = 22.882149 (0.535 sec)\n",
"INFO:tensorflow:loss = 22.882149, step = 770 (0.535 sec)\n",
"INFO:tensorflow:global_step/sec: 1.5433\n",
"INFO:tensorflow:train_accuracy = 0.51753324, train_loss = 29.840134 (0.648 sec)\n",
"INFO:tensorflow:loss = 29.840134, step = 771 (0.648 sec)\n",
"INFO:tensorflow:global_step/sec: 1.97443\n",
"INFO:tensorflow:train_accuracy = 0.52570695, train_loss = 23.693638 (0.507 sec)\n",
"INFO:tensorflow:loss = 23.693638, step = 772 (0.507 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88503\n",
"INFO:tensorflow:train_accuracy = 0.51005024, train_loss = 27.173664 (0.530 sec)\n",
"INFO:tensorflow:loss = 27.173664, step = 773 (0.530 sec)\n",
"INFO:tensorflow:global_step/sec: 1.9211\n",
"INFO:tensorflow:train_accuracy = 0.521164, train_loss = 23.286837 (0.521 sec)\n",
"INFO:tensorflow:loss = 23.286837, step = 774 (0.520 sec)\n",
"INFO:tensorflow:global_step/sec: 1.84103\n",
"INFO:tensorflow:train_accuracy = 0.505618, train_loss = 26.08062 (0.543 sec)\n",
"INFO:tensorflow:loss = 26.08062, step = 775 (0.543 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71688\n",
"INFO:tensorflow:train_accuracy = 0.5198973, train_loss = 26.363838 (0.583 sec)\n",
"INFO:tensorflow:loss = 26.363838, step = 776 (0.583 sec)\n",
"INFO:tensorflow:global_step/sec: 1.96512\n",
"INFO:tensorflow:train_accuracy = 0.54508746, train_loss = 19.960114 (0.509 sec)\n",
"INFO:tensorflow:loss = 19.960114, step = 777 (0.508 sec)\n",
"INFO:tensorflow:global_step/sec: 1.76364\n",
"INFO:tensorflow:train_accuracy = 0.5192308, train_loss = 25.267483 (0.567 sec)\n",
"INFO:tensorflow:loss = 25.267483, step = 778 (0.567 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89967\n",
"INFO:tensorflow:train_accuracy = 0.54435486, train_loss = 21.680595 (0.526 sec)\n",
"INFO:tensorflow:loss = 21.680595, step = 779 (0.526 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91394\n",
"INFO:tensorflow:train_accuracy = 0.5064433, train_loss = 25.967276 (0.522 sec)\n",
"INFO:tensorflow:loss = 25.967276, step = 780 (0.522 sec)\n",
"INFO:tensorflow:global_step/sec: 4.01994\n",
"INFO:tensorflow:train_accuracy = 0.47488585, train_loss = 41.411377 (0.249 sec)\n",
"INFO:tensorflow:loss = 41.411377, step = 781 (0.249 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82776\n",
"INFO:tensorflow:train_accuracy = 0.5343008, train_loss = 21.871437 (0.547 sec)\n",
"INFO:tensorflow:loss = 21.871437, step = 782 (0.547 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61946\n",
"INFO:tensorflow:train_accuracy = 0.50813514, train_loss = 26.583672 (0.618 sec)\n",
"INFO:tensorflow:loss = 26.583672, step = 783 (0.618 sec)\n",
"INFO:tensorflow:global_step/sec: 1.42851\n",
"INFO:tensorflow:train_accuracy = 0.499388, train_loss = 29.173939 (0.700 sec)\n",
"INFO:tensorflow:loss = 29.173939, step = 784 (0.700 sec)\n",
"INFO:tensorflow:global_step/sec: 1.70759\n",
"INFO:tensorflow:train_accuracy = 0.49753696, train_loss = 29.549881 (0.585 sec)\n",
"INFO:tensorflow:loss = 29.549881, step = 785 (0.585 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91402\n",
"INFO:tensorflow:train_accuracy = 0.5260347, train_loss = 22.056356 (0.522 sec)\n",
"INFO:tensorflow:loss = 22.056356, step = 786 (0.523 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73169\n",
"INFO:tensorflow:train_accuracy = 0.52393615, train_loss = 21.680115 (0.577 sec)\n",
"INFO:tensorflow:loss = 21.680115, step = 787 (0.577 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64945\n",
"INFO:tensorflow:train_accuracy = 0.5, train_loss = 27.962715 (0.607 sec)\n",
"INFO:tensorflow:loss = 27.962715, step = 788 (0.607 sec)\n",
"INFO:tensorflow:global_step/sec: 1.70589\n",
"INFO:tensorflow:train_accuracy = 0.5329815, train_loss = 22.635538 (0.586 sec)\n",
"INFO:tensorflow:loss = 22.635538, step = 789 (0.586 sec)\n",
"INFO:tensorflow:global_step/sec: 2.04777\n",
"INFO:tensorflow:train_accuracy = 0.5239948, train_loss = 23.73035 (0.488 sec)\n",
"INFO:tensorflow:loss = 23.73035, step = 790 (0.488 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65284\n",
"INFO:tensorflow:train_accuracy = 0.52680963, train_loss = 21.923597 (0.605 sec)\n",
"INFO:tensorflow:loss = 21.923597, step = 791 (0.605 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63914\n",
"INFO:tensorflow:train_accuracy = 0.53079945, train_loss = 24.881643 (0.610 sec)\n",
"INFO:tensorflow:loss = 24.881643, step = 792 (0.610 sec)\n",
"INFO:tensorflow:global_step/sec: 1.47109\n",
"INFO:tensorflow:train_accuracy = 0.5301205, train_loss = 22.928864 (0.680 sec)\n",
"INFO:tensorflow:loss = 22.928864, step = 793 (0.680 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85488\n",
"INFO:tensorflow:train_accuracy = 0.5133418, train_loss = 25.216295 (0.539 sec)\n",
"INFO:tensorflow:loss = 25.216295, step = 794 (0.539 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88344\n",
"INFO:tensorflow:train_accuracy = 0.49803922, train_loss = 22.482801 (0.531 sec)\n",
"INFO:tensorflow:loss = 22.482801, step = 795 (0.531 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89889\n",
"INFO:tensorflow:train_accuracy = 0.5350993, train_loss = 24.07623 (0.527 sec)\n",
"INFO:tensorflow:loss = 24.07623, step = 796 (0.527 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80269\n",
"INFO:tensorflow:train_accuracy = 0.53018373, train_loss = 21.98014 (0.554 sec)\n",
"INFO:tensorflow:loss = 21.98014, step = 797 (0.554 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71847\n",
"INFO:tensorflow:train_accuracy = 0.52791876, train_loss = 28.297762 (0.582 sec)\n",
"INFO:tensorflow:loss = 28.297762, step = 798 (0.582 sec)\n",
"INFO:tensorflow:Saving checkpoints for 800 into asr-quartznet/model.ckpt.\n",
"INFO:tensorflow:Calling model_fn.\n",
"INFO:tensorflow:Done calling model_fn.\n",
"INFO:tensorflow:Starting evaluation at 2020-10-25T19:03:01Z\n",
"INFO:tensorflow:Graph was finalized.\n",
"INFO:tensorflow:Restoring parameters from asr-quartznet/model.ckpt-800\n",
"INFO:tensorflow:Running local_init_op.\n",
"INFO:tensorflow:Done running local_init_op.\n",
"INFO:tensorflow:Evaluation [1/10]\n",
"INFO:tensorflow:Evaluation [2/10]\n",
"INFO:tensorflow:Evaluation [3/10]\n",
"INFO:tensorflow:Evaluation [4/10]\n",
"INFO:tensorflow:Evaluation [5/10]\n",
"INFO:tensorflow:Evaluation [6/10]\n",
"INFO:tensorflow:Evaluation [7/10]\n",
"INFO:tensorflow:Evaluation [8/10]\n",
"INFO:tensorflow:Evaluation [9/10]\n",
"INFO:tensorflow:Evaluation [10/10]\n",
"INFO:tensorflow:Finished evaluation at 2020-10-25-19:03:10\n",
"INFO:tensorflow:Saving dict for global step 800: WER = 1.0, accuracy = 0.4877301, global_step = 800, loss = 23.982439\n",
"INFO:tensorflow:Saving 'checkpoint_path' summary for global step 800: asr-quartznet/model.ckpt-800\n",
"INFO:tensorflow:global_step/sec: 0.058921\n",
"INFO:tensorflow:train_accuracy = 0.5273698, train_loss = 21.14624 (16.971 sec)\n",
"INFO:tensorflow:loss = 21.14624, step = 799 (16.971 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62908\n",
"INFO:tensorflow:train_accuracy = 0.51856595, train_loss = 25.450104 (0.616 sec)\n",
"INFO:tensorflow:loss = 25.450104, step = 800 (0.616 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82734\n",
"INFO:tensorflow:train_accuracy = 0.5013089, train_loss = 25.118227 (0.546 sec)\n",
"INFO:tensorflow:loss = 25.118227, step = 801 (0.545 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82003\n",
"INFO:tensorflow:train_accuracy = 0.5246114, train_loss = 22.718306 (0.549 sec)\n",
"INFO:tensorflow:loss = 22.718306, step = 802 (0.549 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61895\n",
"INFO:tensorflow:train_accuracy = 0.4866242, train_loss = 26.246677 (0.618 sec)\n",
"INFO:tensorflow:loss = 26.246677, step = 803 (0.618 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75165\n",
"INFO:tensorflow:train_accuracy = 0.5294118, train_loss = 23.632244 (0.571 sec)\n",
"INFO:tensorflow:loss = 23.632244, step = 804 (0.571 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86634\n",
"INFO:tensorflow:train_accuracy = 0.51915455, train_loss = 25.2184 (0.535 sec)\n",
"INFO:tensorflow:loss = 25.2184, step = 805 (0.535 sec)\n",
"INFO:tensorflow:global_step/sec: 1.95629\n",
"INFO:tensorflow:train_accuracy = 0.51856595, train_loss = 23.94997 (0.511 sec)\n",
"INFO:tensorflow:loss = 23.94997, step = 806 (0.511 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91555\n",
"INFO:tensorflow:train_accuracy = 0.5350066, train_loss = 22.399673 (0.522 sec)\n",
"INFO:tensorflow:loss = 22.399673, step = 807 (0.522 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71415\n",
"INFO:tensorflow:train_accuracy = 0.52393275, train_loss = 23.781761 (0.583 sec)\n",
"INFO:tensorflow:loss = 23.781761, step = 808 (0.583 sec)\n",
"INFO:tensorflow:global_step/sec: 1.6268\n",
"INFO:tensorflow:train_accuracy = 0.507732, train_loss = 26.207417 (0.615 sec)\n",
"INFO:tensorflow:loss = 26.207417, step = 809 (0.614 sec)\n",
"INFO:tensorflow:global_step/sec: 1.96122\n",
"INFO:tensorflow:train_accuracy = 0.5103627, train_loss = 23.867794 (0.510 sec)\n",
"INFO:tensorflow:loss = 23.867794, step = 810 (0.510 sec)\n",
"INFO:tensorflow:global_step/sec: 1.53907\n",
"INFO:tensorflow:train_accuracy = 0.49936947, train_loss = 28.069084 (0.650 sec)\n",
"INFO:tensorflow:loss = 28.069084, step = 811 (0.650 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8367\n",
"INFO:tensorflow:train_accuracy = 0.52056557, train_loss = 25.280977 (0.545 sec)\n",
"INFO:tensorflow:loss = 25.280977, step = 812 (0.545 sec)\n",
"INFO:tensorflow:global_step/sec: 1.53521\n",
"INFO:tensorflow:train_accuracy = 0.50062895, train_loss = 27.328949 (0.651 sec)\n",
"INFO:tensorflow:loss = 27.328949, step = 813 (0.651 sec)\n",
"INFO:tensorflow:global_step/sec: 1.6573\n",
"INFO:tensorflow:train_accuracy = 0.5360419, train_loss = 22.734606 (0.603 sec)\n",
"INFO:tensorflow:loss = 22.734606, step = 814 (0.603 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7459\n",
"INFO:tensorflow:train_accuracy = 0.4975669, train_loss = 27.790012 (0.573 sec)\n",
"INFO:tensorflow:loss = 27.790012, step = 815 (0.572 sec)\n",
"INFO:tensorflow:global_step/sec: 1.5966\n",
"INFO:tensorflow:train_accuracy = 0.53040105, train_loss = 24.832186 (0.626 sec)\n",
"INFO:tensorflow:loss = 24.832186, step = 816 (0.626 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71405\n",
"INFO:tensorflow:train_accuracy = 0.51772153, train_loss = 24.533976 (0.584 sec)\n",
"INFO:tensorflow:loss = 24.533976, step = 817 (0.584 sec)\n",
"INFO:tensorflow:global_step/sec: 1.81796\n",
"INFO:tensorflow:train_accuracy = 0.49936306, train_loss = 25.702198 (0.550 sec)\n",
"INFO:tensorflow:loss = 25.702198, step = 818 (0.550 sec)\n",
"INFO:tensorflow:global_step/sec: 1.54164\n",
"INFO:tensorflow:train_accuracy = 0.53359175, train_loss = 25.729666 (0.649 sec)\n",
"INFO:tensorflow:loss = 25.729666, step = 819 (0.649 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69908\n",
"INFO:tensorflow:train_accuracy = 0.52618134, train_loss = 25.70068 (0.588 sec)\n",
"INFO:tensorflow:loss = 25.70068, step = 820 (0.588 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86754\n",
"INFO:tensorflow:train_accuracy = 0.51856595, train_loss = 24.065329 (0.536 sec)\n",
"INFO:tensorflow:loss = 24.065329, step = 821 (0.536 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71321\n",
"INFO:tensorflow:train_accuracy = 0.47994986, train_loss = 26.935024 (0.583 sec)\n",
"INFO:tensorflow:loss = 26.935024, step = 822 (0.584 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75993\n",
"INFO:tensorflow:train_accuracy = 0.52248675, train_loss = 23.211458 (0.568 sec)\n",
"INFO:tensorflow:loss = 23.211458, step = 823 (0.568 sec)\n",
"INFO:tensorflow:global_step/sec: 1.81021\n",
"INFO:tensorflow:train_accuracy = 0.49747476, train_loss = 26.24081 (0.552 sec)\n",
"INFO:tensorflow:loss = 26.24081, step = 824 (0.552 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91678\n",
"INFO:tensorflow:train_accuracy = 0.5272727, train_loss = 23.338968 (0.522 sec)\n",
"INFO:tensorflow:loss = 23.338968, step = 825 (0.523 sec)\n",
"INFO:tensorflow:global_step/sec: 1.9605\n",
"INFO:tensorflow:train_accuracy = 0.5487805, train_loss = 20.555069 (0.510 sec)\n",
"INFO:tensorflow:loss = 20.555069, step = 826 (0.510 sec)\n",
"INFO:tensorflow:global_step/sec: 5.75507\n",
"INFO:tensorflow:train_accuracy = 0.49302325, train_loss = 24.794947 (0.174 sec)\n",
"INFO:tensorflow:loss = 24.794947, step = 827 (0.174 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69513\n",
"INFO:tensorflow:train_accuracy = 0.52597404, train_loss = 24.243855 (0.590 sec)\n",
"INFO:tensorflow:loss = 24.243855, step = 828 (0.589 sec)\n",
"INFO:tensorflow:global_step/sec: 1.58764\n",
"INFO:tensorflow:train_accuracy = 0.5006353, train_loss = 24.575933 (0.630 sec)\n",
"INFO:tensorflow:loss = 24.575933, step = 829 (0.630 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66412\n",
"INFO:tensorflow:train_accuracy = 0.5192308, train_loss = 24.14315 (0.601 sec)\n",
"INFO:tensorflow:loss = 24.14315, step = 830 (0.601 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74861\n",
"INFO:tensorflow:train_accuracy = 0.46764347, train_loss = 29.602222 (0.572 sec)\n",
"INFO:tensorflow:loss = 29.602222, step = 831 (0.572 sec)\n",
"INFO:tensorflow:global_step/sec: 1.44295\n",
"INFO:tensorflow:train_accuracy = 0.5081556, train_loss = 31.028538 (0.693 sec)\n",
"INFO:tensorflow:loss = 31.028538, step = 832 (0.693 sec)\n",
"INFO:tensorflow:global_step/sec: 2.0295\n",
"INFO:tensorflow:train_accuracy = 0.5420561, train_loss = 22.346298 (0.493 sec)\n",
"INFO:tensorflow:loss = 22.346298, step = 833 (0.493 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72141\n",
"INFO:tensorflow:train_accuracy = 0.532, train_loss = 22.729603 (0.581 sec)\n",
"INFO:tensorflow:loss = 22.729603, step = 834 (0.581 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72341\n",
"INFO:tensorflow:train_accuracy = 0.51856595, train_loss = 23.210796 (0.580 sec)\n",
"INFO:tensorflow:loss = 23.210796, step = 835 (0.580 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62817\n",
"INFO:tensorflow:train_accuracy = 0.51450187, train_loss = 24.98116 (0.623 sec)\n",
"INFO:tensorflow:loss = 24.98116, step = 836 (0.622 sec)\n",
"INFO:tensorflow:global_step/sec: 1.42283\n",
"INFO:tensorflow:train_accuracy = 0.5310436, train_loss = 22.760098 (0.694 sec)\n",
"INFO:tensorflow:loss = 22.760098, step = 837 (0.694 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85797\n",
"INFO:tensorflow:train_accuracy = 0.54435486, train_loss = 20.98231 (0.538 sec)\n",
"INFO:tensorflow:loss = 20.98231, step = 838 (0.538 sec)\n",
"INFO:tensorflow:global_step/sec: 1.67675\n",
"INFO:tensorflow:train_accuracy = 0.5343008, train_loss = 23.53356 (0.596 sec)\n",
"INFO:tensorflow:loss = 23.53356, step = 839 (0.596 sec)\n",
"INFO:tensorflow:global_step/sec: 1.80051\n",
"INFO:tensorflow:train_accuracy = 0.5335968, train_loss = 23.0472 (0.556 sec)\n",
"INFO:tensorflow:loss = 23.0472, step = 840 (0.556 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65122\n",
"INFO:tensorflow:train_accuracy = 0.51005024, train_loss = 26.182693 (0.606 sec)\n",
"INFO:tensorflow:loss = 26.182693, step = 841 (0.606 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89536\n",
"INFO:tensorflow:train_accuracy = 0.5413333, train_loss = 22.153236 (0.527 sec)\n",
"INFO:tensorflow:loss = 22.153236, step = 842 (0.527 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82537\n",
"INFO:tensorflow:train_accuracy = 0.53079945, train_loss = 22.814346 (0.548 sec)\n",
"INFO:tensorflow:loss = 22.814346, step = 843 (0.548 sec)\n",
"INFO:tensorflow:global_step/sec: 1.95759\n",
"INFO:tensorflow:train_accuracy = 0.5252918, train_loss = 24.09735 (0.511 sec)\n",
"INFO:tensorflow:loss = 24.09735, step = 844 (0.510 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86884\n",
"INFO:tensorflow:train_accuracy = 0.5413333, train_loss = 22.079529 (0.535 sec)\n",
"INFO:tensorflow:loss = 22.079529, step = 845 (0.536 sec)\n",
"INFO:tensorflow:global_step/sec: 1.84322\n",
"INFO:tensorflow:train_accuracy = 0.52393275, train_loss = 23.448343 (0.542 sec)\n",
"INFO:tensorflow:loss = 23.448343, step = 846 (0.543 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88149\n",
"INFO:tensorflow:train_accuracy = 0.5349144, train_loss = 22.740362 (0.531 sec)\n",
"INFO:tensorflow:loss = 22.740362, step = 847 (0.531 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61571\n",
"INFO:tensorflow:train_accuracy = 0.50318474, train_loss = 26.456608 (0.619 sec)\n",
"INFO:tensorflow:loss = 26.456608, step = 848 (0.619 sec)\n",
"INFO:tensorflow:global_step/sec: 1.84199\n",
"INFO:tensorflow:train_accuracy = 0.5191816, train_loss = 24.333862 (0.543 sec)\n",
"INFO:tensorflow:loss = 24.333862, step = 849 (0.543 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77186\n",
"INFO:tensorflow:train_accuracy = 0.5089744, train_loss = 23.870316 (0.564 sec)\n",
"INFO:tensorflow:loss = 23.870316, step = 850 (0.564 sec)\n",
"INFO:tensorflow:global_step/sec: 2.01192\n",
"INFO:tensorflow:train_accuracy = 0.5418327, train_loss = 24.544678 (0.497 sec)\n",
"INFO:tensorflow:loss = 24.544678, step = 851 (0.497 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8588\n",
"INFO:tensorflow:train_accuracy = 0.51846963, train_loss = 24.528423 (0.538 sec)\n",
"INFO:tensorflow:loss = 24.528423, step = 852 (0.538 sec)\n",
"INFO:tensorflow:global_step/sec: 1.95511\n",
"INFO:tensorflow:train_accuracy = 0.50319284, train_loss = 23.478653 (0.511 sec)\n",
"INFO:tensorflow:loss = 23.478653, step = 853 (0.511 sec)\n",
"INFO:tensorflow:global_step/sec: 1.68706\n",
"INFO:tensorflow:train_accuracy = 0.52893674, train_loss = 22.103542 (0.593 sec)\n",
"INFO:tensorflow:loss = 22.103542, step = 854 (0.593 sec)\n",
"INFO:tensorflow:global_step/sec: 1.92492\n",
"INFO:tensorflow:train_accuracy = 0.54072094, train_loss = 22.865028 (0.520 sec)\n",
"INFO:tensorflow:loss = 22.865028, step = 855 (0.520 sec)\n",
"INFO:tensorflow:global_step/sec: 1.53862\n",
"INFO:tensorflow:train_accuracy = 0.4856115, train_loss = 31.275969 (0.650 sec)\n",
"INFO:tensorflow:loss = 31.275969, step = 856 (0.649 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64082\n",
"INFO:tensorflow:train_accuracy = 0.53737116, train_loss = 23.949528 (0.609 sec)\n",
"INFO:tensorflow:loss = 23.949528, step = 857 (0.609 sec)\n",
"INFO:tensorflow:global_step/sec: 1.56093\n",
"INFO:tensorflow:train_accuracy = 0.49625936, train_loss = 32.62814 (0.641 sec)\n",
"INFO:tensorflow:loss = 32.62814, step = 858 (0.641 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88905\n",
"INFO:tensorflow:train_accuracy = 0.51846963, train_loss = 23.519531 (0.530 sec)\n",
"INFO:tensorflow:loss = 23.519531, step = 859 (0.530 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64827\n",
"INFO:tensorflow:train_accuracy = 0.52734375, train_loss = 22.838219 (0.606 sec)\n",
"INFO:tensorflow:loss = 22.838219, step = 860 (0.606 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72578\n",
"INFO:tensorflow:train_accuracy = 0.500615, train_loss = 27.567211 (0.579 sec)\n",
"INFO:tensorflow:loss = 27.567211, step = 861 (0.579 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75984\n",
"INFO:tensorflow:train_accuracy = 0.53157896, train_loss = 28.914532 (0.568 sec)\n",
"INFO:tensorflow:loss = 28.914532, step = 862 (0.568 sec)\n",
"INFO:tensorflow:global_step/sec: 1.68237\n",
"INFO:tensorflow:train_accuracy = 0.4885404, train_loss = 28.839916 (0.595 sec)\n",
"INFO:tensorflow:loss = 28.839916, step = 863 (0.595 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74001\n",
"INFO:tensorflow:train_accuracy = 0.513308, train_loss = 26.64303 (0.575 sec)\n",
"INFO:tensorflow:loss = 26.64303, step = 864 (0.575 sec)\n",
"INFO:tensorflow:global_step/sec: 1.9186\n",
"INFO:tensorflow:train_accuracy = 0.51432294, train_loss = 28.674835 (0.521 sec)\n",
"INFO:tensorflow:loss = 28.674835, step = 865 (0.521 sec)\n",
"INFO:tensorflow:global_step/sec: 1.54814\n",
"INFO:tensorflow:train_accuracy = 0.47715735, train_loss = 28.696083 (0.646 sec)\n",
"INFO:tensorflow:loss = 28.696083, step = 866 (0.646 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88998\n",
"INFO:tensorflow:train_accuracy = 0.5103896, train_loss = 31.769606 (0.529 sec)\n",
"INFO:tensorflow:loss = 31.769606, step = 867 (0.529 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88463\n",
"INFO:tensorflow:train_accuracy = 0.53149605, train_loss = 23.699232 (0.531 sec)\n",
"INFO:tensorflow:loss = 23.699232, step = 868 (0.531 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77183\n",
"INFO:tensorflow:train_accuracy = 0.49557522, train_loss = 31.683483 (0.564 sec)\n",
"INFO:tensorflow:loss = 31.683483, step = 869 (0.563 sec)\n",
"INFO:tensorflow:global_step/sec: 1.83292\n",
"INFO:tensorflow:train_accuracy = 0.5232558, train_loss = 25.781986 (0.546 sec)\n",
"INFO:tensorflow:loss = 25.781986, step = 870 (0.547 sec)\n",
"INFO:tensorflow:global_step/sec: 1.6926\n",
"INFO:tensorflow:train_accuracy = 0.46599495, train_loss = 35.175503 (0.591 sec)\n",
"INFO:tensorflow:loss = 35.175503, step = 871 (0.590 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90436\n",
"INFO:tensorflow:train_accuracy = 0.5458221, train_loss = 22.491962 (0.525 sec)\n",
"INFO:tensorflow:loss = 22.491962, step = 872 (0.526 sec)\n",
"INFO:tensorflow:global_step/sec: 5.60812\n",
"INFO:tensorflow:train_accuracy = 0.53, train_loss = 23.977997 (0.178 sec)\n",
"INFO:tensorflow:loss = 23.977997, step = 873 (0.178 sec)\n",
"INFO:tensorflow:global_step/sec: 1.696\n",
"INFO:tensorflow:train_accuracy = 0.49674055, train_loss = 23.871347 (0.590 sec)\n",
"INFO:tensorflow:loss = 23.871347, step = 874 (0.589 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61426\n",
"INFO:tensorflow:train_accuracy = 0.5, train_loss = 31.317385 (0.620 sec)\n",
"INFO:tensorflow:loss = 31.317385, step = 875 (0.620 sec)\n",
"INFO:tensorflow:global_step/sec: 1.41081\n",
"INFO:tensorflow:train_accuracy = 0.45387453, train_loss = 35.517204 (0.709 sec)\n",
"INFO:tensorflow:loss = 35.517204, step = 876 (0.708 sec)\n",
"INFO:tensorflow:global_step/sec: 1.83209\n",
"INFO:tensorflow:train_accuracy = 0.48908857, train_loss = 30.526398 (0.546 sec)\n",
"INFO:tensorflow:loss = 30.526398, step = 877 (0.546 sec)\n",
"INFO:tensorflow:global_step/sec: 1.6371\n",
"INFO:tensorflow:train_accuracy = 0.4906367, train_loss = 31.039234 (0.611 sec)\n",
"INFO:tensorflow:loss = 31.039234, step = 878 (0.611 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91581\n",
"INFO:tensorflow:train_accuracy = 0.494133, train_loss = 27.866758 (0.522 sec)\n",
"INFO:tensorflow:loss = 27.866758, step = 879 (0.521 sec)\n",
"INFO:tensorflow:global_step/sec: 1.69624\n",
"INFO:tensorflow:train_accuracy = 0.45945945, train_loss = 34.782524 (0.590 sec)\n",
"INFO:tensorflow:loss = 34.782524, step = 880 (0.590 sec)\n",
"INFO:tensorflow:global_step/sec: 1.75317\n",
"INFO:tensorflow:train_accuracy = 0.47682118, train_loss = 28.840279 (0.570 sec)\n",
"INFO:tensorflow:loss = 28.840279, step = 881 (0.570 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61861\n",
"INFO:tensorflow:train_accuracy = 0.47916666, train_loss = 28.579689 (0.620 sec)\n",
"INFO:tensorflow:loss = 28.579689, step = 882 (0.620 sec)\n",
"INFO:tensorflow:global_step/sec: 1.43195\n",
"INFO:tensorflow:train_accuracy = 0.43121693, train_loss = 29.560726 (0.697 sec)\n",
"INFO:tensorflow:loss = 29.560726, step = 883 (0.697 sec)\n",
"INFO:tensorflow:global_step/sec: 1.94682\n",
"INFO:tensorflow:train_accuracy = 0.5045752, train_loss = 27.993214 (0.514 sec)\n",
"INFO:tensorflow:loss = 27.993214, step = 884 (0.514 sec)\n",
"INFO:tensorflow:global_step/sec: 1.94521\n",
"INFO:tensorflow:train_accuracy = 0.48598132, train_loss = 25.31596 (0.514 sec)\n",
"INFO:tensorflow:loss = 25.31596, step = 885 (0.514 sec)\n",
"INFO:tensorflow:global_step/sec: 1.98873\n",
"INFO:tensorflow:train_accuracy = 0.5051151, train_loss = 25.7466 (0.503 sec)\n",
"INFO:tensorflow:loss = 25.7466, step = 886 (0.502 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90777\n",
"INFO:tensorflow:train_accuracy = 0.49737534, train_loss = 28.146023 (0.524 sec)\n",
"INFO:tensorflow:loss = 28.146023, step = 887 (0.524 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66928\n",
"INFO:tensorflow:train_accuracy = 0.52380955, train_loss = 24.852644 (0.600 sec)\n",
"INFO:tensorflow:loss = 24.852644, step = 888 (0.600 sec)\n",
"INFO:tensorflow:global_step/sec: 2.01381\n",
"INFO:tensorflow:train_accuracy = 0.5176471, train_loss = 24.228374 (0.496 sec)\n",
"INFO:tensorflow:loss = 24.228374, step = 889 (0.496 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82249\n",
"INFO:tensorflow:train_accuracy = 0.4897436, train_loss = 30.799898 (0.549 sec)\n",
"INFO:tensorflow:loss = 30.799898, step = 890 (0.549 sec)\n",
"INFO:tensorflow:global_step/sec: 1.82461\n",
"INFO:tensorflow:train_accuracy = 0.5162549, train_loss = 25.190605 (0.548 sec)\n",
"INFO:tensorflow:loss = 25.190605, step = 891 (0.549 sec)\n",
"INFO:tensorflow:global_step/sec: 1.94875\n",
"INFO:tensorflow:train_accuracy = 0.4867021, train_loss = 28.898216 (0.513 sec)\n",
"INFO:tensorflow:loss = 28.898216, step = 892 (0.513 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64231\n",
"INFO:tensorflow:train_accuracy = 0.4934896, train_loss = 29.726278 (0.608 sec)\n",
"INFO:tensorflow:loss = 29.726278, step = 893 (0.608 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64013\n",
"INFO:tensorflow:train_accuracy = 0.496134, train_loss = 29.29231 (0.610 sec)\n",
"INFO:tensorflow:loss = 29.29231, step = 894 (0.610 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8499\n",
"INFO:tensorflow:train_accuracy = 0.49167734, train_loss = 28.575726 (0.541 sec)\n",
"INFO:tensorflow:loss = 28.575726, step = 895 (0.541 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77797\n",
"INFO:tensorflow:train_accuracy = 0.47493404, train_loss = 28.053143 (0.562 sec)\n",
"INFO:tensorflow:loss = 28.053143, step = 896 (0.562 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86205\n",
"INFO:tensorflow:train_accuracy = 0.48416886, train_loss = 27.203428 (0.537 sec)\n",
"INFO:tensorflow:loss = 27.203428, step = 897 (0.538 sec)\n",
"INFO:tensorflow:global_step/sec: 1.98181\n",
"INFO:tensorflow:train_accuracy = 0.49271524, train_loss = 26.616747 (0.504 sec)\n",
"INFO:tensorflow:loss = 26.616747, step = 898 (0.504 sec)\n",
"INFO:tensorflow:global_step/sec: 1.81731\n",
"INFO:tensorflow:train_accuracy = 0.47058824, train_loss = 32.199 (0.550 sec)\n",
"INFO:tensorflow:loss = 32.199, step = 899 (0.550 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7335\n",
"INFO:tensorflow:train_accuracy = 0.46560848, train_loss = 28.141626 (0.578 sec)\n",
"INFO:tensorflow:loss = 28.141626, step = 900 (0.578 sec)\n",
"INFO:tensorflow:global_step/sec: 1.54674\n",
"INFO:tensorflow:train_accuracy = 0.37752524, train_loss = 34.612434 (0.646 sec)\n",
"INFO:tensorflow:loss = 34.612434, step = 901 (0.646 sec)\n",
"INFO:tensorflow:global_step/sec: 1.85395\n",
"INFO:tensorflow:train_accuracy = 0.45614034, train_loss = 31.57251 (0.539 sec)\n",
"INFO:tensorflow:loss = 31.57251, step = 902 (0.539 sec)\n",
"INFO:tensorflow:global_step/sec: 2.00452\n",
"INFO:tensorflow:train_accuracy = 0.48152867, train_loss = 27.885271 (0.499 sec)\n",
"INFO:tensorflow:loss = 27.885271, step = 903 (0.499 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88747\n",
"INFO:tensorflow:train_accuracy = 0.4901704, train_loss = 27.515907 (0.530 sec)\n",
"INFO:tensorflow:loss = 27.515907, step = 904 (0.530 sec)\n",
"INFO:tensorflow:global_step/sec: 1.93617\n",
"INFO:tensorflow:train_accuracy = 0.43320364, train_loss = 30.552681 (0.516 sec)\n",
"INFO:tensorflow:loss = 30.552681, step = 905 (0.516 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66202\n",
"INFO:tensorflow:train_accuracy = 0.48341838, train_loss = 29.74511 (0.602 sec)\n",
"INFO:tensorflow:loss = 29.74511, step = 906 (0.602 sec)\n",
"INFO:tensorflow:global_step/sec: 1.76148\n",
"INFO:tensorflow:train_accuracy = 0.46375, train_loss = 29.977573 (0.568 sec)\n",
"INFO:tensorflow:loss = 29.977573, step = 907 (0.567 sec)\n",
"INFO:tensorflow:global_step/sec: 1.54126\n",
"INFO:tensorflow:train_accuracy = 0.43467012, train_loss = 30.844913 (0.649 sec)\n",
"INFO:tensorflow:loss = 30.844913, step = 908 (0.649 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64478\n",
"INFO:tensorflow:train_accuracy = 0.4741488, train_loss = 31.178621 (0.608 sec)\n",
"INFO:tensorflow:loss = 31.178621, step = 909 (0.608 sec)\n",
"INFO:tensorflow:global_step/sec: 1.55353\n",
"INFO:tensorflow:train_accuracy = 0.41269842, train_loss = 33.379894 (0.644 sec)\n",
"INFO:tensorflow:loss = 33.379894, step = 910 (0.644 sec)\n",
"INFO:tensorflow:global_step/sec: 1.99289\n",
"INFO:tensorflow:train_accuracy = 0.4465409, train_loss = 31.468464 (0.502 sec)\n",
"INFO:tensorflow:loss = 31.468464, step = 911 (0.502 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90132\n",
"INFO:tensorflow:train_accuracy = 0.40673575, train_loss = 27.39006 (0.525 sec)\n",
"INFO:tensorflow:loss = 27.39006, step = 912 (0.525 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74134\n",
"INFO:tensorflow:train_accuracy = 0.46183205, train_loss = 30.020407 (0.575 sec)\n",
"INFO:tensorflow:loss = 30.020407, step = 913 (0.575 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8492\n",
"INFO:tensorflow:train_accuracy = 0.4744898, train_loss = 28.398506 (0.541 sec)\n",
"INFO:tensorflow:loss = 28.398506, step = 914 (0.541 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74684\n",
"INFO:tensorflow:train_accuracy = 0.43076923, train_loss = 29.346638 (0.572 sec)\n",
"INFO:tensorflow:loss = 29.346638, step = 915 (0.572 sec)\n",
"INFO:tensorflow:global_step/sec: 1.9137\n",
"INFO:tensorflow:train_accuracy = 0.41666666, train_loss = 28.040676 (0.522 sec)\n",
"INFO:tensorflow:loss = 28.040676, step = 916 (0.522 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89982\n",
"INFO:tensorflow:train_accuracy = 0.47003993, train_loss = 27.63496 (0.527 sec)\n",
"INFO:tensorflow:loss = 27.63496, step = 917 (0.527 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90216\n",
"INFO:tensorflow:train_accuracy = 0.46920052, train_loss = 27.308065 (0.526 sec)\n",
"INFO:tensorflow:loss = 27.308065, step = 918 (0.526 sec)\n",
"INFO:tensorflow:global_step/sec: 6.30036\n",
"INFO:tensorflow:train_accuracy = 0.40191388, train_loss = 25.713858 (0.159 sec)\n",
"INFO:tensorflow:loss = 25.713858, step = 919 (0.158 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71791\n",
"INFO:tensorflow:train_accuracy = 0.4443038, train_loss = 29.161219 (0.582 sec)\n",
"INFO:tensorflow:loss = 29.161219, step = 920 (0.582 sec)\n",
"INFO:tensorflow:global_step/sec: 1.62437\n",
"INFO:tensorflow:train_accuracy = 0.4224464, train_loss = 30.629715 (0.615 sec)\n",
"INFO:tensorflow:loss = 30.629715, step = 921 (0.615 sec)\n",
"INFO:tensorflow:global_step/sec: 1.4476\n",
"INFO:tensorflow:train_accuracy = 0.40276036, train_loss = 32.740864 (0.691 sec)\n",
"INFO:tensorflow:loss = 32.740864, step = 922 (0.692 sec)\n",
"INFO:tensorflow:global_step/sec: 1.84381\n",
"INFO:tensorflow:train_accuracy = 0.47493404, train_loss = 31.524172 (0.542 sec)\n",
"INFO:tensorflow:loss = 31.524172, step = 923 (0.542 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64604\n",
"INFO:tensorflow:train_accuracy = 0.47191012, train_loss = 29.738382 (0.608 sec)\n",
"INFO:tensorflow:loss = 29.738382, step = 924 (0.608 sec)\n",
"INFO:tensorflow:global_step/sec: 1.67625\n",
"INFO:tensorflow:train_accuracy = 0.4654498, train_loss = 27.428383 (0.596 sec)\n",
"INFO:tensorflow:loss = 27.428383, step = 925 (0.596 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88461\n",
"INFO:tensorflow:train_accuracy = 0.47129506, train_loss = 29.250671 (0.530 sec)\n",
"INFO:tensorflow:loss = 29.250671, step = 926 (0.530 sec)\n",
"INFO:tensorflow:global_step/sec: 1.67987\n",
"INFO:tensorflow:train_accuracy = 0.42729592, train_loss = 29.471851 (0.596 sec)\n",
"INFO:tensorflow:loss = 29.471851, step = 927 (0.596 sec)\n",
"INFO:tensorflow:global_step/sec: 1.60898\n",
"INFO:tensorflow:train_accuracy = 0.47651008, train_loss = 25.750168 (0.622 sec)\n",
"INFO:tensorflow:loss = 25.750168, step = 928 (0.622 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91774\n",
"INFO:tensorflow:train_accuracy = 0.46184212, train_loss = 29.771572 (0.521 sec)\n",
"INFO:tensorflow:loss = 29.771572, step = 929 (0.521 sec)\n",
"INFO:tensorflow:global_step/sec: 1.45981\n",
"INFO:tensorflow:train_accuracy = 0.4644243, train_loss = 26.730326 (0.685 sec)\n",
"INFO:tensorflow:loss = 26.730326, step = 930 (0.685 sec)\n",
"INFO:tensorflow:global_step/sec: 1.815\n",
"INFO:tensorflow:train_accuracy = 0.48963732, train_loss = 29.314228 (0.551 sec)\n",
"INFO:tensorflow:loss = 29.314228, step = 931 (0.551 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8851\n",
"INFO:tensorflow:train_accuracy = 0.5085414, train_loss = 27.056107 (0.530 sec)\n",
"INFO:tensorflow:loss = 27.056107, step = 932 (0.530 sec)\n",
"INFO:tensorflow:global_step/sec: 2.01621\n",
"INFO:tensorflow:train_accuracy = 0.45714286, train_loss = 28.40454 (0.496 sec)\n",
"INFO:tensorflow:loss = 28.40454, step = 933 (0.496 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86713\n",
"INFO:tensorflow:train_accuracy = 0.5201072, train_loss = 23.763395 (0.536 sec)\n",
"INFO:tensorflow:loss = 23.763395, step = 934 (0.536 sec)\n",
"INFO:tensorflow:global_step/sec: 1.81193\n",
"INFO:tensorflow:train_accuracy = 0.5044586, train_loss = 27.943136 (0.552 sec)\n",
"INFO:tensorflow:loss = 27.943136, step = 935 (0.552 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88526\n",
"INFO:tensorflow:train_accuracy = 0.51295334, train_loss = 27.23473 (0.530 sec)\n",
"INFO:tensorflow:loss = 27.23473, step = 936 (0.530 sec)\n",
"INFO:tensorflow:global_step/sec: 2.04103\n",
"INFO:tensorflow:train_accuracy = 0.5185676, train_loss = 24.011147 (0.490 sec)\n",
"INFO:tensorflow:loss = 24.011147, step = 937 (0.489 sec)\n",
"INFO:tensorflow:global_step/sec: 2.03189\n",
"INFO:tensorflow:train_accuracy = 0.49935484, train_loss = 26.892612 (0.492 sec)\n",
"INFO:tensorflow:loss = 26.892612, step = 938 (0.493 sec)\n",
"INFO:tensorflow:global_step/sec: 1.83598\n",
"INFO:tensorflow:train_accuracy = 0.5181818, train_loss = 27.918762 (0.545 sec)\n",
"INFO:tensorflow:loss = 27.918762, step = 939 (0.545 sec)\n",
"INFO:tensorflow:global_step/sec: 1.76731\n",
"INFO:tensorflow:train_accuracy = 0.49159122, train_loss = 26.735216 (0.566 sec)\n",
"INFO:tensorflow:loss = 26.735216, step = 940 (0.565 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63913\n",
"INFO:tensorflow:train_accuracy = 0.48358586, train_loss = 28.95419 (0.610 sec)\n",
"INFO:tensorflow:loss = 28.95419, step = 941 (0.611 sec)\n",
"INFO:tensorflow:global_step/sec: 1.79294\n",
"INFO:tensorflow:train_accuracy = 0.48677248, train_loss = 25.70279 (0.558 sec)\n",
"INFO:tensorflow:loss = 25.70279, step = 942 (0.557 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86671\n",
"INFO:tensorflow:train_accuracy = 0.5185185, train_loss = 24.408476 (0.536 sec)\n",
"INFO:tensorflow:loss = 24.408476, step = 943 (0.536 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86003\n",
"INFO:tensorflow:train_accuracy = 0.5109114, train_loss = 26.476772 (0.537 sec)\n",
"INFO:tensorflow:loss = 26.476772, step = 944 (0.537 sec)\n",
"INFO:tensorflow:global_step/sec: 1.73968\n",
"INFO:tensorflow:train_accuracy = 0.52476573, train_loss = 25.745953 (0.575 sec)\n",
"INFO:tensorflow:loss = 25.745953, step = 945 (0.575 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72682\n",
"INFO:tensorflow:train_accuracy = 0.49216712, train_loss = 27.185413 (0.579 sec)\n",
"INFO:tensorflow:loss = 27.185413, step = 946 (0.578 sec)\n",
"INFO:tensorflow:global_step/sec: 1.5398\n",
"INFO:tensorflow:train_accuracy = 0.46827412, train_loss = 29.999231 (0.650 sec)\n",
"INFO:tensorflow:loss = 29.999231, step = 947 (0.650 sec)\n",
"INFO:tensorflow:global_step/sec: 1.92507\n",
"INFO:tensorflow:train_accuracy = 0.48284626, train_loss = 29.611286 (0.519 sec)\n",
"INFO:tensorflow:loss = 29.611286, step = 948 (0.519 sec)\n",
"INFO:tensorflow:global_step/sec: 1.7525\n",
"INFO:tensorflow:train_accuracy = 0.49051833, train_loss = 28.212967 (0.571 sec)\n",
"INFO:tensorflow:loss = 28.212967, step = 949 (0.571 sec)\n",
"INFO:tensorflow:global_step/sec: 1.95376\n",
"INFO:tensorflow:train_accuracy = 0.49935484, train_loss = 24.357634 (0.512 sec)\n",
"INFO:tensorflow:loss = 24.357634, step = 950 (0.512 sec)\n",
"INFO:tensorflow:global_step/sec: 1.52489\n",
"INFO:tensorflow:train_accuracy = 0.4853129, train_loss = 27.964537 (0.656 sec)\n",
"INFO:tensorflow:loss = 27.964537, step = 951 (0.656 sec)\n",
"INFO:tensorflow:global_step/sec: 1.98443\n",
"INFO:tensorflow:train_accuracy = 0.5153846, train_loss = 26.167154 (0.504 sec)\n",
"INFO:tensorflow:loss = 26.167154, step = 952 (0.504 sec)\n",
"INFO:tensorflow:global_step/sec: 1.74661\n",
"INFO:tensorflow:train_accuracy = 0.45925927, train_loss = 30.698273 (0.573 sec)\n",
"INFO:tensorflow:loss = 30.698273, step = 953 (0.573 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88429\n",
"INFO:tensorflow:train_accuracy = 0.51462764, train_loss = 23.909204 (0.530 sec)\n",
"INFO:tensorflow:loss = 23.909204, step = 954 (0.530 sec)\n",
"INFO:tensorflow:global_step/sec: 1.55753\n",
"INFO:tensorflow:train_accuracy = 0.4723618, train_loss = 28.367178 (0.642 sec)\n",
"INFO:tensorflow:loss = 28.367178, step = 955 (0.643 sec)\n",
"INFO:tensorflow:global_step/sec: 1.67044\n",
"INFO:tensorflow:train_accuracy = 0.47820672, train_loss = 30.399902 (0.599 sec)\n",
"INFO:tensorflow:loss = 30.399902, step = 956 (0.599 sec)\n",
"INFO:tensorflow:global_step/sec: 1.68831\n",
"INFO:tensorflow:train_accuracy = 0.5119798, train_loss = 26.849676 (0.592 sec)\n",
"INFO:tensorflow:loss = 26.849676, step = 957 (0.592 sec)\n",
"INFO:tensorflow:global_step/sec: 1.90055\n",
"INFO:tensorflow:train_accuracy = 0.52879584, train_loss = 24.290825 (0.526 sec)\n",
"INFO:tensorflow:loss = 24.290825, step = 958 (0.526 sec)\n",
"INFO:tensorflow:global_step/sec: 1.70791\n",
"INFO:tensorflow:train_accuracy = 0.44458437, train_loss = 30.908335 (0.585 sec)\n",
"INFO:tensorflow:loss = 30.908335, step = 959 (0.586 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77447\n",
"INFO:tensorflow:train_accuracy = 0.46683672, train_loss = 31.552433 (0.563 sec)\n",
"INFO:tensorflow:loss = 31.552433, step = 960 (0.563 sec)\n",
"INFO:tensorflow:global_step/sec: 2.03389\n",
"INFO:tensorflow:train_accuracy = 0.4812903, train_loss = 25.807554 (0.492 sec)\n",
"INFO:tensorflow:loss = 25.807554, step = 961 (0.492 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88177\n",
"INFO:tensorflow:train_accuracy = 0.5218509, train_loss = 23.460094 (0.531 sec)\n",
"INFO:tensorflow:loss = 23.460094, step = 962 (0.531 sec)\n",
"INFO:tensorflow:global_step/sec: 1.99576\n",
"INFO:tensorflow:train_accuracy = 0.5354223, train_loss = 24.167751 (0.501 sec)\n",
"INFO:tensorflow:loss = 24.167751, step = 963 (0.501 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89098\n",
"INFO:tensorflow:train_accuracy = 0.5368421, train_loss = 23.083664 (0.529 sec)\n",
"INFO:tensorflow:loss = 23.083664, step = 964 (0.529 sec)\n",
"INFO:tensorflow:global_step/sec: 5.78483\n",
"INFO:tensorflow:train_accuracy = 0.51141554, train_loss = 29.763493 (0.173 sec)\n",
"INFO:tensorflow:loss = 29.763493, step = 965 (0.172 sec)\n",
"INFO:tensorflow:global_step/sec: 1.68977\n",
"INFO:tensorflow:train_accuracy = 0.4812903, train_loss = 27.267563 (0.592 sec)\n",
"INFO:tensorflow:loss = 27.267563, step = 966 (0.592 sec)\n",
"INFO:tensorflow:global_step/sec: 1.66399\n",
"INFO:tensorflow:train_accuracy = 0.5130208, train_loss = 25.057673 (0.601 sec)\n",
"INFO:tensorflow:loss = 25.057673, step = 967 (0.601 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77202\n",
"INFO:tensorflow:train_accuracy = 0.49240506, train_loss = 30.594067 (0.564 sec)\n",
"INFO:tensorflow:loss = 30.594067, step = 968 (0.564 sec)\n",
"INFO:tensorflow:global_step/sec: 1.41851\n",
"INFO:tensorflow:train_accuracy = 0.49939686, train_loss = 30.429632 (0.705 sec)\n",
"INFO:tensorflow:loss = 30.429632, step = 969 (0.705 sec)\n",
"INFO:tensorflow:global_step/sec: 1.91322\n",
"INFO:tensorflow:train_accuracy = 0.49615383, train_loss = 25.959751 (0.523 sec)\n",
"INFO:tensorflow:loss = 25.959751, step = 970 (0.523 sec)\n",
"INFO:tensorflow:global_step/sec: 1.61966\n",
"INFO:tensorflow:train_accuracy = 0.5096277, train_loss = 25.458817 (0.617 sec)\n",
"INFO:tensorflow:loss = 25.458817, step = 971 (0.616 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71108\n",
"INFO:tensorflow:train_accuracy = 0.5208614, train_loss = 22.842987 (0.585 sec)\n",
"INFO:tensorflow:loss = 22.842987, step = 972 (0.586 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88667\n",
"INFO:tensorflow:train_accuracy = 0.51111114, train_loss = 25.011816 (0.530 sec)\n",
"INFO:tensorflow:loss = 25.011816, step = 973 (0.530 sec)\n",
"INFO:tensorflow:global_step/sec: 1.613\n",
"INFO:tensorflow:train_accuracy = 0.51472473, train_loss = 24.89143 (0.620 sec)\n",
"INFO:tensorflow:loss = 24.89143, step = 974 (0.620 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89672\n",
"INFO:tensorflow:train_accuracy = 0.5531629, train_loss = 21.915455 (0.527 sec)\n",
"INFO:tensorflow:loss = 21.915455, step = 975 (0.527 sec)\n",
"INFO:tensorflow:global_step/sec: 1.45474\n",
"INFO:tensorflow:train_accuracy = 0.52213544, train_loss = 25.846222 (0.688 sec)\n",
"INFO:tensorflow:loss = 25.846222, step = 976 (0.688 sec)\n",
"INFO:tensorflow:global_step/sec: 1.65326\n",
"INFO:tensorflow:train_accuracy = 0.5248042, train_loss = 24.240417 (0.605 sec)\n",
"INFO:tensorflow:loss = 24.240417, step = 977 (0.605 sec)\n",
"INFO:tensorflow:global_step/sec: 1.86049\n",
"INFO:tensorflow:train_accuracy = 0.5032595, train_loss = 25.828854 (0.537 sec)\n",
"INFO:tensorflow:loss = 25.828854, step = 978 (0.537 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8213\n",
"INFO:tensorflow:train_accuracy = 0.5124183, train_loss = 26.725288 (0.549 sec)\n",
"INFO:tensorflow:loss = 26.725288, step = 979 (0.550 sec)\n",
"INFO:tensorflow:global_step/sec: 1.94313\n",
"INFO:tensorflow:train_accuracy = 0.530343, train_loss = 23.661865 (0.514 sec)\n",
"INFO:tensorflow:loss = 23.661865, step = 980 (0.513 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8251\n",
"INFO:tensorflow:train_accuracy = 0.502551, train_loss = 25.46909 (0.548 sec)\n",
"INFO:tensorflow:loss = 25.46909, step = 981 (0.549 sec)\n",
"INFO:tensorflow:global_step/sec: 1.89149\n",
"INFO:tensorflow:train_accuracy = 0.52140075, train_loss = 24.38026 (0.529 sec)\n",
"INFO:tensorflow:loss = 24.38026, step = 982 (0.529 sec)\n",
"INFO:tensorflow:global_step/sec: 2.12975\n",
"INFO:tensorflow:train_accuracy = 0.53002614, train_loss = 22.987545 (0.469 sec)\n",
"INFO:tensorflow:loss = 22.987545, step = 983 (0.469 sec)\n",
"INFO:tensorflow:global_step/sec: 1.84003\n",
"INFO:tensorflow:train_accuracy = 0.5489418, train_loss = 22.585402 (0.543 sec)\n",
"INFO:tensorflow:loss = 22.585402, step = 984 (0.543 sec)\n",
"INFO:tensorflow:global_step/sec: 1.81985\n",
"INFO:tensorflow:train_accuracy = 0.5325365, train_loss = 23.645592 (0.550 sec)\n",
"INFO:tensorflow:loss = 23.645592, step = 985 (0.550 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88226\n",
"INFO:tensorflow:train_accuracy = 0.5038071, train_loss = 26.304844 (0.531 sec)\n",
"INFO:tensorflow:loss = 26.304844, step = 986 (0.531 sec)\n",
"INFO:tensorflow:global_step/sec: 1.6338\n",
"INFO:tensorflow:train_accuracy = 0.51225805, train_loss = 25.26618 (0.612 sec)\n",
"INFO:tensorflow:loss = 25.26618, step = 987 (0.613 sec)\n",
"INFO:tensorflow:global_step/sec: 1.76678\n",
"INFO:tensorflow:train_accuracy = 0.5228162, train_loss = 22.909328 (0.566 sec)\n",
"INFO:tensorflow:loss = 22.909328, step = 988 (0.566 sec)\n",
"INFO:tensorflow:global_step/sec: 1.64154\n",
"INFO:tensorflow:train_accuracy = 0.5181818, train_loss = 25.061558 (0.609 sec)\n",
"INFO:tensorflow:loss = 25.061558, step = 989 (0.609 sec)\n",
"INFO:tensorflow:global_step/sec: 1.6837\n",
"INFO:tensorflow:train_accuracy = 0.5411765, train_loss = 25.154573 (0.594 sec)\n",
"INFO:tensorflow:loss = 25.154573, step = 990 (0.594 sec)\n",
"INFO:tensorflow:global_step/sec: 1.8817\n",
"INFO:tensorflow:train_accuracy = 0.5498008, train_loss = 22.334309 (0.532 sec)\n",
"INFO:tensorflow:loss = 22.334309, step = 991 (0.532 sec)\n",
"INFO:tensorflow:global_step/sec: 1.71527\n",
"INFO:tensorflow:train_accuracy = 0.51428574, train_loss = 25.280369 (0.583 sec)\n",
"INFO:tensorflow:loss = 25.280369, step = 992 (0.582 sec)\n",
"INFO:tensorflow:global_step/sec: 1.53405\n",
"INFO:tensorflow:train_accuracy = 0.5096525, train_loss = 29.859425 (0.652 sec)\n",
"INFO:tensorflow:loss = 29.859425, step = 993 (0.651 sec)\n",
"INFO:tensorflow:global_step/sec: 1.63333\n",
"INFO:tensorflow:train_accuracy = 0.52611464, train_loss = 25.629574 (0.613 sec)\n",
"INFO:tensorflow:loss = 25.629574, step = 994 (0.613 sec)\n",
"INFO:tensorflow:global_step/sec: 1.88513\n",
"INFO:tensorflow:train_accuracy = 0.5167095, train_loss = 26.12327 (0.530 sec)\n",
"INFO:tensorflow:loss = 26.12327, step = 995 (0.530 sec)\n",
"INFO:tensorflow:global_step/sec: 1.72887\n",
"INFO:tensorflow:train_accuracy = 0.51969504, train_loss = 26.131132 (0.579 sec)\n",
"INFO:tensorflow:loss = 26.131132, step = 996 (0.579 sec)\n",
"INFO:tensorflow:global_step/sec: 1.49556\n",
"INFO:tensorflow:train_accuracy = 0.5144291, train_loss = 28.200312 (0.668 sec)\n",
"INFO:tensorflow:loss = 28.200312, step = 997 (0.668 sec)\n",
"INFO:tensorflow:global_step/sec: 1.77259\n",
"INFO:tensorflow:train_accuracy = 0.51808786, train_loss = 24.605595 (0.564 sec)\n",
"INFO:tensorflow:loss = 24.605595, step = 998 (0.564 sec)\n",
"INFO:tensorflow:Saving checkpoints for 1000 into asr-quartznet/model.ckpt.\n",
"WARNING:tensorflow:From /home/husein/.local/lib/python3.6/site-packages/tensorflow_core/python/training/saver.py:963: remove_checkpoint (from tensorflow.python.training.checkpoint_management) is deprecated and will be removed in a future version.\n",
"Instructions for updating:\n",
"Use standard file APIs to delete files with this prefix.\n",
"INFO:tensorflow:Calling model_fn.\n",
"INFO:tensorflow:Done calling model_fn.\n",
"INFO:tensorflow:Starting evaluation at 2020-10-25T19:05:11Z\n",
"INFO:tensorflow:Graph was finalized.\n",
"INFO:tensorflow:Restoring parameters from asr-quartznet/model.ckpt-1000\n",
"INFO:tensorflow:Running local_init_op.\n",
"INFO:tensorflow:Done running local_init_op.\n",
"INFO:tensorflow:Evaluation [1/10]\n",
"INFO:tensorflow:Evaluation [2/10]\n",
"INFO:tensorflow:Evaluation [3/10]\n",
"INFO:tensorflow:Evaluation [4/10]\n",
"INFO:tensorflow:Evaluation [5/10]\n",
"INFO:tensorflow:Evaluation [6/10]\n",
"INFO:tensorflow:Evaluation [7/10]\n",
"INFO:tensorflow:Evaluation [8/10]\n",
"INFO:tensorflow:Evaluation [9/10]\n",
"INFO:tensorflow:Evaluation [10/10]\n",
"INFO:tensorflow:Finished evaluation at 2020-10-25-19:05:19\n",
"INFO:tensorflow:Saving dict for global step 1000: WER = 1.0, accuracy = 0.5030675, global_step = 1000, loss = 24.014952\n",
"INFO:tensorflow:Saving 'checkpoint_path' summary for global step 1000: asr-quartznet/model.ckpt-1000\n",
"INFO:tensorflow:global_step/sec: 0.0570441\n",
"INFO:tensorflow:train_accuracy = 0.516129, train_loss = 26.32674 (17.530 sec)\n",
"INFO:tensorflow:loss = 26.32674, step = 999 (17.531 sec)\n",
"INFO:tensorflow:Loss for final step: 26.32674.\n"
]
}
],
"source": [
"train.run_training(\n",
" train_fn = train_dataset,\n",
" model_fn = model_fn,\n",
" model_dir = 'asr-quartznet',\n",
" num_gpus = 1,\n",
" log_step = 1,\n",
" save_checkpoint_step = 200,\n",
" max_steps = train_steps,\n",
" eval_fn = dev_dataset,\n",
" train_hooks = train_hooks,\n",
")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Reference\n",
"\n",
"1. Nvidia OpenSeq2Seq, https://github.com/NVIDIA/OpenSeq2Seq/blob/master/example_configs/speech2text/quartznet15x5_LibriSpeech.py"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.7"
}
},
"nbformat": 4,
"nbformat_minor": 2
}