17_XX_DockerExperiment.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from misc.preferences import PREFERENCES\n",
    "from misc.run_configuration import hyperOpt_goodParams, default_params\n",
    "from misc.experimental_environment import Experiment\n",
    "import time\n",
    "from data.organic2019 import organic_dataset as dsl"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "PREFERENCES.defaults(\n",
    "    data_root='./data/data/organic2019',\n",
    "    data_train='train.csv',    \n",
    "    data_validation='validation.csv',\n",
    "    data_test='test.csv',\n",
    "    source_index=0,\n",
    "    target_vocab_index=1,\n",
    "    file_format='csv'\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Experiment DockerTest initialized\n",
      "START: 1555516902.6268125\n",
      "Log path is  /data/home/felix/ABSA-Transformer/logs/DockerTest/20190417/16\n",
      "Log path is  /data/home/felix/ABSA-Transformer/logs/DockerTest/20190417/17\n",
      "\n",
      "\n",
      "#########################################################################\n",
      "Name: DockerTest\n",
      "Description: [DOCKER_OFF] This experiment is run to measure the performance decrease when using docker vs. running it locally.\n",
      "#########################################################################\n",
      "\n",
      "\n",
      "+----------------------------------------------------------------------------------+\n",
      "|                                 Hyperparameters                                  |\n",
      "+------------------------------+---------------------------------------------------+\n",
      "|          Parameter           |                       Value                       |\n",
      "+------------------------------+---------------------------------------------------+\n",
      "|            kwargs            | {'batch_size': 12, 'learning_rate_schedu[...]one} |\n",
      "|          model_size          |                        300                        |\n",
      "|        early_stopping        |                         5                         |\n",
      "|           use_cuda           |                        True                       |\n",
      "|             task             |                      entities                     |\n",
      "|          batch_size          |                         12                        |\n",
      "| learning_rate_scheduler_type |             LearningSchedulerType.Noam            |\n",
      "|      output_layer_type       |             OutputLayerType.LinearSum             |\n",
      "|        optimizer_type        |                 OptimizerType.Adam                |\n",
      "|        learning_rate         |                      7.2e-05                      |\n",
      "|  noam_learning_rate_warmup   |                        8000                       |\n",
      "|  noam_learning_rate_factor   |                       1.418                       |\n",
      "|          adam_beta1          |                        0.81                       |\n",
      "|          adam_beta2          |                       0.7173                      |\n",
      "|           adam_eps           |                      0.000814                     |\n",
      "|      adam_weight_decay       |                         0                         |\n",
      "|         adam_amsgrad         |                       False                       |\n",
      "|           use_bias           |                       False                       |\n",
      "|         n_enc_blocks         |                         2                         |\n",
      "|           n_heads            |                         1                         |\n",
      "|             d_k              |                        300                        |\n",
      "|             d_v              |                        300                        |\n",
      "|         dropout_rate         |                      0.302424                     |\n",
      "|     pointwise_layer_size     |                        405                        |\n",
      "|      last_layer_dropout      |                  0.79602089766246                 |\n",
      "|   log_every_xth_iteration    |                         -1                        |\n",
      "|          num_epochs          |                         5                         |\n",
      "|        embedding_type        |                       glove                       |\n",
      "|        embedding_name        |                         6B                        |\n",
      "|        embedding_dim         |                        300                        |\n",
      "|       clip_comments_to       |                        113                        |\n",
      "|      finetune_embedding      |                        True                       |\n",
      "|           language           |                         en                        |\n",
      "|        use_stop_words        |                        True                       |\n",
      "|         use_stemming         |                       False                       |\n",
      "|        harmonize_bahn        |                        True                       |\n",
      "|      use_spell_checkers      |                       False                       |\n",
      "|      replace_url_tokens      |                        True                       |\n",
      "|       use_text_cleaner       |                       False                       |\n",
      "|     contraction_removal      |                       False                       |\n",
      "|    organic_text_cleaning     |                       False                       |\n",
      "|             seed             |                        None                       |\n",
      "+------------------------------+---------------------------------------------------+\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "                                                                    \r"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "+------------------------+\n",
      "| GERM EVAL 2017 DATASET |\n",
      "+---------------+--------+\n",
      "|     Split     |  Size  |\n",
      "+---------------+--------+\n",
      "|     train     |  8441  |\n",
      "|   validation  |  745   |\n",
      "|      test     |  694   |\n",
      "+---------------+--------+\n",
      "+------------------------------------------+\n",
      "|             Vocabulary Stats             |\n",
      "+----------------------------------+-------+\n",
      "|            Vocabulary            |  Size |\n",
      "+----------------------------------+-------+\n",
      "|        aspect_sentiments         |   4   |\n",
      "|             comments             | 11685 |\n",
      "|             padding              |   2   |\n",
      "| GMOs genetic engineering general |   4   |\n",
      "|      conventional companies      |   3   |\n",
      "|       conventional farming       |   4   |\n",
      "|       conventional general       |   4   |\n",
      "|      conventional products       |   4   |\n",
      "|        organic companies         |   4   |\n",
      "|         organic farmers          |   4   |\n",
      "|         organic general          |   4   |\n",
      "|         organic products         |   4   |\n",
      "+----------------------------------+-------+\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "+-------------------------------------------------------------------+\n",
      "|                  GMOs genetic engineering general                 |\n",
      "+-------------+---------+--------------------+----------------------+\n",
      "|    Label    | Samples |   Triv. Accuracy   |     Class Weight     |\n",
      "+-------------+---------+--------------------+----------------------+\n",
      "|     n/a     |   9606  | 97.22672064777328  | 0.027732793522267207 |\n",
      "|   negative  |    87   | 0.8805668016194331 |  0.9911943319838057  |\n",
      "|   neutral   |   135   | 1.3663967611336032 |  0.986336032388664   |\n",
      "|   positive  |    52   | 0.5263157894736842 |  0.9947368421052631  |\n",
      "|     Sum     |   9880  |                    |         1.0          |\n",
      "| Head Weight |         |                    |  0.9722672064777328  |\n",
      "+-------------+---------+--------------------+----------------------+\n",
      "\n",
      "\n",
      "\n",
      "+--------------------------------------------------------------------+\n",
      "|                       conventional companies                       |\n",
      "+-------------+---------+---------------------+----------------------+\n",
      "|    Label    | Samples |    Triv. Accuracy   |     Class Weight     |\n",
      "+-------------+---------+---------------------+----------------------+\n",
      "|     n/a     |   9858  |  99.77732793522267  | 0.002226720647773295 |\n",
      "|   neutral   |    10   | 0.10121457489878542 |  0.9989878542510121  |\n",
      "|   negative  |    12   | 0.12145748987854252 |  0.9987854251012146  |\n",
      "|     Sum     |   9880  |                     |         1.0          |\n",
      "| Head Weight |         |                     |  0.9977732793522267  |\n",
      "+-------------+---------+---------------------+----------------------+\n",
      "\n",
      "\n",
      "\n",
      "+-------------------------------------------------------------------+\n",
      "|                        conventional farming                       |\n",
      "+-------------+---------+--------------------+----------------------+\n",
      "|    Label    | Samples |   Triv. Accuracy   |     Class Weight     |\n",
      "+-------------+---------+--------------------+----------------------+\n",
      "|     n/a     |   9644  | 97.61133603238866  | 0.023886639676113375 |\n",
      "|   neutral   |    82   | 0.8299595141700404 |  0.9917004048582996  |\n",
      "|   negative  |   114   | 1.153846153846154  |  0.9884615384615385  |\n",
      "|   positive  |    40   | 0.4048582995951417 |  0.9959514170040485  |\n",
      "|     Sum     |   9880  |                    |         1.0          |\n",
      "| Head Weight |         |                    |  0.9761133603238866  |\n",
      "+-------------+---------+--------------------+----------------------+\n",
      "\n",
      "\n",
      "\n",
      "+--------------------------------------------------------------------+\n",
      "|                        conventional general                        |\n",
      "+-------------+---------+---------------------+----------------------+\n",
      "|    Label    | Samples |    Triv. Accuracy   |     Class Weight     |\n",
      "+-------------+---------+---------------------+----------------------+\n",
      "|     n/a     |   9833  |   99.5242914979757  | 0.004757085020242968 |\n",
      "|   positive  |    7    |  0.0708502024291498 |  0.9992914979757085  |\n",
      "|   negative  |    17   | 0.17206477732793524 |  0.9982793522267206  |\n",
      "|   neutral   |    23   |  0.2327935222672065 |  0.997672064777328   |\n",
      "|     Sum     |   9880  |                     |         1.0          |\n",
      "| Head Weight |         |                     |  0.995242914979757   |\n",
      "+-------------+---------+---------------------+----------------------+\n",
      "\n",
      "\n",
      "\n",
      "+-------------------------------------------------------------------+\n",
      "|                       conventional products                       |\n",
      "+-------------+---------+--------------------+----------------------+\n",
      "|    Label    | Samples |   Triv. Accuracy   |     Class Weight     |\n",
      "+-------------+---------+--------------------+----------------------+\n",
      "|     n/a     |   9564  | 96.80161943319838  | 0.031983805668016174 |\n",
      "|   negative  |   145   | 1.4676113360323886 |  0.9853238866396761  |\n",
      "|   positive  |    65   | 0.6578947368421052 |  0.993421052631579   |\n",
      "|   neutral   |   106   | 1.0728744939271255 |  0.9892712550607288  |\n",
      "|     Sum     |   9880  |                    |         1.0          |\n",
      "| Head Weight |         |                    |  0.9680161943319838  |\n",
      "+-------------+---------+--------------------+----------------------+\n",
      "\n",
      "\n",
      "\n",
      "+-------------------------------------------------------------------+\n",
      "|                         organic companies                         |\n",
      "+-------------+---------+--------------------+----------------------+\n",
      "|    Label    | Samples |   Triv. Accuracy   |     Class Weight     |\n",
      "+-------------+---------+--------------------+----------------------+\n",
      "|     n/a     |   9554  | 96.70040485829959  | 0.032995951417004066 |\n",
      "|   positive  |   115   | 1.1639676113360324 |  0.9883603238866396  |\n",
      "|   negative  |    86   | 0.8704453441295547 |  0.9912955465587044  |\n",
      "|   neutral   |   125   | 1.2651821862348178 |  0.9873481781376519  |\n",
      "|     Sum     |   9880  |                    |         1.0          |\n",
      "| Head Weight |         |                    |  0.9670040485829959  |\n",
      "+-------------+---------+--------------------+----------------------+\n",
      "\n",
      "\n",
      "\n",
      "+-----------------------------------------------------------------+\n",
      "|                         organic farmers                         |\n",
      "+-------------+---------+-------------------+---------------------+\n",
      "|    Label    | Samples |   Triv. Accuracy  |     Class Weight    |\n",
      "+-------------+---------+-------------------+---------------------+\n",
      "|     n/a     |   8922  | 90.30364372469636 | 0.09696356275303641 |\n",
      "|   positive  |   311   | 3.147773279352227 |  0.9685222672064777 |\n",
      "|   negative  |   240   |  2.42914979757085 |  0.9757085020242915 |\n",
      "|   neutral   |   407   | 4.119433198380567 |  0.9588056680161944 |\n",
      "|     Sum     |   9880  |                   |         1.0         |\n",
      "| Head Weight |         |                   |  0.9030364372469636 |\n",
      "+-------------+---------+-------------------+---------------------+\n",
      "\n",
      "\n",
      "\n",
      "+------------------------------------------------------------------+\n",
      "|                         organic general                          |\n",
      "+-------------+---------+--------------------+---------------------+\n",
      "|    Label    | Samples |   Triv. Accuracy   |     Class Weight    |\n",
      "+-------------+---------+--------------------+---------------------+\n",
      "|   neutral   |   659   | 6.670040485829959  |  0.9332995951417005 |\n",
      "|     n/a     |   8660  | 87.65182186234817  | 0.12348178137651822 |\n",
      "|   positive  |   286   | 2.8947368421052633 |  0.9710526315789474 |\n",
      "|   negative  |   275   | 2.783400809716599  |  0.972165991902834  |\n",
      "|     Sum     |   9880  |                    |         1.0         |\n",
      "| Head Weight |         |                    |  0.8765182186234818 |\n",
      "+-------------+---------+--------------------+---------------------+\n",
      "\n",
      "\n",
      "\n",
      "+----------------------------------------------------------------+\n",
      "|                        organic products                        |\n",
      "+-------------+---------+-------------------+--------------------+\n",
      "|    Label    | Samples |   Triv. Accuracy  |    Class Weight    |\n",
      "+-------------+---------+-------------------+--------------------+\n",
      "|     n/a     |   8053  |  81.5080971659919 | 0.184919028340081  |\n",
      "|   positive  |   745   | 7.540485829959515 | 0.9245951417004048 |\n",
      "|   negative  |   504   | 5.101214574898785 | 0.9489878542510122 |\n",
      "|   neutral   |   578   | 5.850202429149798 | 0.941497975708502  |\n",
      "|     Sum     |   9880  |                   |        1.0         |\n",
      "| Head Weight |         |                   | 0.815080971659919  |\n",
      "+-------------+---------+-------------------+--------------------+\n",
      "\n",
      "\n",
      "+--------------------------------------------------------------+\n",
      "|                            Train                             |\n",
      "+----------------------------------+------+------+------+------+\n",
      "|             Category             | POS  | NEG  | NEU  | Sum  |\n",
      "+----------------------------------+------+------+------+------+\n",
      "|         organic general          | 258  | 254  | 582  | 1094 |\n",
      "|      conventional companies      |  0   |  12  |  8   |  20  |\n",
      "|         organic products         | 715  | 475  | 496  | 1686 |\n",
      "|        organic companies         | 100  |  80  | 108  | 288  |\n",
      "|       conventional general       |  6   |  16  |  20  |  42  |\n",
      "|         organic farmers          | 281  | 212  | 360  | 853  |\n",
      "|       conventional farming       |  38  | 113  |  80  | 231  |\n",
      "|      conventional products       |  62  | 151  |  96  | 309  |\n",
      "| GMOs genetic engineering general |  53  |  85  | 117  | 255  |\n",
      "|              Total               | 1513 | 1398 | 1867 | 4778 |\n",
      "+----------------------------------+------+------+------+------+\n",
      "\n",
      "+----------------------------------------------------------+\n",
      "|                        Validation                        |\n",
      "+----------------------------------+-----+-----+-----+-----+\n",
      "|             Category             | POS | NEG | NEU | Sum |\n",
      "+----------------------------------+-----+-----+-----+-----+\n",
      "|       conventional general       |  1  |  0  |  1  |  2  |\n",
      "| GMOs genetic engineering general |  2  |  2  |  4  |  8  |\n",
      "|      conventional products       |  4  |  7  |  4  |  15 |\n",
      "|         organic general          |  14 |  13 |  45 |  72 |\n",
      "|        organic companies         |  19 |  10 |  14 |  43 |\n",
      "|         organic products         |  69 |  40 |  50 | 159 |\n",
      "|         organic farmers          |  17 |  19 |  28 |  64 |\n",
      "|       conventional farming       |  5  |  6  |  0  |  11 |\n",
      "|      conventional companies      |  0  |  1  |  0  |  1  |\n",
      "|              Total               | 131 |  98 | 146 | 375 |\n",
      "+----------------------------------+-----+-----+-----+-----+\n",
      "\n",
      "+----------------------------------------------------------+\n",
      "|                           Test                           |\n",
      "+----------------------------------+-----+-----+-----+-----+\n",
      "|             Category             | POS | NEG | NEU | Sum |\n",
      "+----------------------------------+-----+-----+-----+-----+\n",
      "|         organic general          |  35 |  24 |  44 | 103 |\n",
      "|      conventional products       |  6  |  9  |  7  |  22 |\n",
      "|         organic products         |  49 |  37 |  48 | 134 |\n",
      "|         organic farmers          |  44 |  18 |  23 |  85 |\n",
      "| GMOs genetic engineering general |  3  |  7  |  15 |  25 |\n",
      "|       conventional general       |  0  |  1  |  2  |  3  |\n",
      "|       conventional farming       |  0  |  5  |  3  |  8  |\n",
      "|      conventional companies      |  0  |  0  |  2  |  2  |\n",
      "|        organic companies         |  10 |  6  |  8  |  24 |\n",
      "|              Total               | 147 | 107 | 152 | 406 |\n",
      "+----------------------------------+-----+-----+-----+-----+\n",
      "\n",
      "\n",
      "dataset loaded. Duration: 1.954179048538208\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "pre_training - INFO - Classes: ['n/a', 'neutral', 'positive', 'negative']\n",
      "pre_training - INFO - Classes: ['n/a', 'neutral', 'positive', 'negative']\n",
      "pre_training - INFO - JointAspectTagger (\n",
      "  (encoder): TransformerEncoder(\n",
      "    (positional_encoding): PositionalEncoding2(\n",
      "      (dropout): Dropout(p=0.302424)\n",
      "    )\n",
      "    (encoder_blocks): ModuleList(\n",
      "      (0): EncoderBlock(\n",
      "        (self_attention_layer): MultiHeadedSelfAttentionLayer(\n",
      "          (query_projections): Linear(in_features=300, out_features=300, bias=False)\n",
      "          (key_projections): Linear(in_features=300, out_features=300, bias=False)\n",
      "          (value_projections): Linear(in_features=300, out_features=300, bias=False)\n",
      "          (attention_layer): ScaledDotProductAttentionLayer(\n",
      "            (dropout): Dropout(p=0.302424)\n",
      "          )\n",
      "          (w_0): Linear(in_features=300, out_features=300, bias=False)\n",
      "          (dropout): Dropout(p=0.302424)\n",
      "        )\n",
      "        (feed_forward_layer): PointWiseFCLayer(\n",
      "          (w_1): Linear(in_features=300, out_features=405, bias=False)\n",
      "          (w_2): Linear(in_features=405, out_features=300, bias=False)\n",
      "          (dropout): Dropout(p=0.302424)\n",
      "        )\n",
      "        (layer_norm): LayerNorm()\n",
      "      )\n",
      "      (1): EncoderBlock(\n",
      "        (self_attention_layer): MultiHeadedSelfAttentionLayer(\n",
      "          (query_projections): Linear(in_features=300, out_features=300, bias=False)\n",
      "          (key_projections): Linear(in_features=300, out_features=300, bias=False)\n",
      "          (value_projections): Linear(in_features=300, out_features=300, bias=False)\n",
      "          (attention_layer): ScaledDotProductAttentionLayer(\n",
      "            (dropout): Dropout(p=0.302424)\n",
      "          )\n",
      "          (w_0): Linear(in_features=300, out_features=300, bias=False)\n",
      "          (dropout): Dropout(p=0.302424)\n",
      "        )\n",
      "        (feed_forward_layer): PointWiseFCLayer(\n",
      "          (w_1): Linear(in_features=300, out_features=405, bias=False)\n",
      "          (w_2): Linear(in_features=405, out_features=300, bias=False)\n",
      "          (dropout): Dropout(p=0.302424)\n",
      "        )\n",
      "        (layer_norm): LayerNorm()\n",
      "      )\n",
      "    )\n",
      "    (layer_norm): LayerNorm()\n",
      "    (src_embeddings): Embedding(11685, 300)\n",
      "  ), weights=((300, 300), (300, 300), (300, 300), (300, 300), (405, 300), (300, 405), (300,), (300,), (300, 300), (300, 300), (300, 300), (300, 300), (405, 300), (300, 405), (300,), (300,), (300,), (300,), (11685, 300)), parameters=4713300\n",
      "  (taggers): ModuleList(\n",
      "    (0): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "    (1): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "    (2): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "    (3): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "    (4): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "    (5): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "    (6): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "    (7): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "    (8): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "  ), weights=((4, 300), (4,), (4, 300), (4,), (4, 300), (4,), (4, 300), (4,), (4, 300), (4,), (4, 300), (4,), (4, 300), (4,), (4, 300), (4,), (4, 300), (4,)), parameters=10836\n",
      ")\n",
      "==================================\n",
      "Total Number of parameters: 4,724,136\n",
      "==================================\n",
      "\n",
      "pre_training - INFO - JointAspectTagger (\n",
      "  (encoder): TransformerEncoder(\n",
      "    (positional_encoding): PositionalEncoding2(\n",
      "      (dropout): Dropout(p=0.302424)\n",
      "    )\n",
      "    (encoder_blocks): ModuleList(\n",
      "      (0): EncoderBlock(\n",
      "        (self_attention_layer): MultiHeadedSelfAttentionLayer(\n",
      "          (query_projections): Linear(in_features=300, out_features=300, bias=False)\n",
      "          (key_projections): Linear(in_features=300, out_features=300, bias=False)\n",
      "          (value_projections): Linear(in_features=300, out_features=300, bias=False)\n",
      "          (attention_layer): ScaledDotProductAttentionLayer(\n",
      "            (dropout): Dropout(p=0.302424)\n",
      "          )\n",
      "          (w_0): Linear(in_features=300, out_features=300, bias=False)\n",
      "          (dropout): Dropout(p=0.302424)\n",
      "        )\n",
      "        (feed_forward_layer): PointWiseFCLayer(\n",
      "          (w_1): Linear(in_features=300, out_features=405, bias=False)\n",
      "          (w_2): Linear(in_features=405, out_features=300, bias=False)\n",
      "          (dropout): Dropout(p=0.302424)\n",
      "        )\n",
      "        (layer_norm): LayerNorm()\n",
      "      )\n",
      "      (1): EncoderBlock(\n",
      "        (self_attention_layer): MultiHeadedSelfAttentionLayer(\n",
      "          (query_projections): Linear(in_features=300, out_features=300, bias=False)\n",
      "          (key_projections): Linear(in_features=300, out_features=300, bias=False)\n",
      "          (value_projections): Linear(in_features=300, out_features=300, bias=False)\n",
      "          (attention_layer): ScaledDotProductAttentionLayer(\n",
      "            (dropout): Dropout(p=0.302424)\n",
      "          )\n",
      "          (w_0): Linear(in_features=300, out_features=300, bias=False)\n",
      "          (dropout): Dropout(p=0.302424)\n",
      "        )\n",
      "        (feed_forward_layer): PointWiseFCLayer(\n",
      "          (w_1): Linear(in_features=300, out_features=405, bias=False)\n",
      "          (w_2): Linear(in_features=405, out_features=300, bias=False)\n",
      "          (dropout): Dropout(p=0.302424)\n",
      "        )\n",
      "        (layer_norm): LayerNorm()\n",
      "      )\n",
      "    )\n",
      "    (layer_norm): LayerNorm()\n",
      "    (src_embeddings): Embedding(11685, 300)\n",
      "  ), weights=((300, 300), (300, 300), (300, 300), (300, 300), (405, 300), (300, 405), (300,), (300,), (300, 300), (300, 300), (300, 300), (300, 300), (405, 300), (300, 405), (300,), (300,), (300,), (300,), (11685, 300)), parameters=4713300\n",
      "  (taggers): ModuleList(\n",
      "    (0): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "    (1): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "    (2): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "    (3): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "    (4): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "    (5): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "    (6): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "    (7): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "    (8): CommentWiseSumLogSoftmax(\n",
      "      (output_projection): Linear(in_features=300, out_features=4, bias=True)\n",
      "      (dropout): Dropout(p=0.79602089766246)\n",
      "    )\n",
      "  ), weights=((4, 300), (4,), (4, 300), (4,), (4, 300), (4,), (4, 300), (4,), (4, 300), (4,), (4, 300), (4,), (4, 300), (4,), (4, 300), (4,), (4, 300), (4,)), parameters=10836\n",
      ")\n",
      "==================================\n",
      "Total Number of parameters: 4,724,136\n",
      "==================================\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "pre_training - INFO - Tensorboard enabled. Run will be located at /runs/DockerTest/20190417/17/20190417/5/0/. Full path is /data/home/felix/ABSA-Transformer/runs/DockerTest/20190417/17/20190417/5EP/0\n",
      "pre_training - INFO - Tensorboard enabled. Run will be located at /runs/DockerTest/20190417/17/20190417/5/0/. Full path is /data/home/felix/ABSA-Transformer/runs/DockerTest/20190417/17/20190417/5EP/0\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/data/anaconda/envs/py35/lib/python3.6/site-packages/torch/onnx/utils.py:501: UserWarning: ONNX export failed on ATen operator std because torch.onnx.symbolic.std does not exist\n",
      "  .format(op_name, op_name))\n",
      "/data/anaconda/envs/py35/lib/python3.6/site-packages/torch/onnx/utils.py:501: UserWarning: ONNX export failed on ATen operator std because torch.onnx.symbolic.std does not exist\n",
      "  .format(op_name, op_name))\n",
      "/data/anaconda/envs/py35/lib/python3.6/site-packages/torch/onnx/utils.py:501: UserWarning: ONNX export failed on ATen operator std because torch.onnx.symbolic.std does not exist\n",
      "  .format(op_name, op_name))\n",
      "/data/anaconda/envs/py35/lib/python3.6/site-packages/torch/onnx/utils.py:501: UserWarning: ONNX export failed on ATen operator std because torch.onnx.symbolic.std does not exist\n",
      "  .format(op_name, op_name))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "pre_training - DEBUG - train with cuda support\n",
      "pre_training - DEBUG - train with cuda support\n",
      "pre_training - INFO - 704 Iterations per epoch with batch size of 12\n",
      "pre_training - INFO - 704 Iterations per epoch with batch size of 12\n",
      "pre_training - INFO - Total iterations: 3520\n",
      "pre_training - INFO - Total iterations: 3520\n",
      "pre_training - INFO - Total number of samples: 42240\n",
      "pre_training - INFO - Total number of samples: 42240\n",
      "pre_training - INFO - START training.\n",
      "pre_training - INFO - START training.\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "55192a2d4c1c4bc98a62f0ee05c8718a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=704), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# EP\t# IT\ttr loss\t\tval loss\tf1\t\tacc\t\tduration / total time\n",
      "1\t8k\t0.63\t\t0.43\t\t0.132\t\t0.888\t\t0.54m - 0.5m / 0.0m\n",
      "2\t17k\t0.53\t\t0.43\t\t0.157\t\t0.847\t\t0.54m - 1.1m / 2.8m\n",
      "3\t25k\t0.50\t\t0.38\t\t0.179\t\t0.887\t\t0.53m - 1.7m / 2.8m\n",
      "4\t34k\t0.49\t\t0.45\t\t0.154\t\t0.843\t\t0.54m - 2.3m / 2.8m\n",
      "5\t42k\t0.48\t\t0.41\t\t0.191\t\t0.886\t\t0.54m - 2.8m / 2.8m\n",
      "Training duration was 172.8538167476654\n",
      "pre_training - DEBUG - --- Valid Scores ---\n",
      "pre_training - DEBUG - --- Valid Scores ---\n",
      "VAL f1\t0.19074074074074074 - (0.19074074074074074)\n",
      "VAL loss\t0.37986882275374473\n",
      "       .---.\n",
      " \t/     \\\n",
      "\t\\.@-@./\t\tExperiment: [0/1]\n",
      "\t/`\\_/`\\\t\tStatus: ok\n",
      "\t//  _  \\\\\tLoss: 0.37986882275374473\n",
      "\t| \\     )|_\tf1: 0.19074074074074074\n",
      "\t/`\\_`>  <_/ \\\n",
      "\t\\__/'---'\\__/\n",
      "\n",
      "#################################################################################\n",
      "############################## EXPERIMENT COMPLETE ##############################\n",
      "\n",
      "\n",
      "Run [0/1]: 0.19482758620689655\n",
      "------------------------------\n",
      "Mean: 0.19482758620689655\n",
      "TEST F1 Statistics\n",
      "count    1.000000\n",
      "mean     0.194828\n",
      "std           NaN\n",
      "min      0.194828\n",
      "25%      0.194828\n",
      "50%      0.194828\n",
      "75%      0.194828\n",
      "max      0.194828\n",
      "Name: test_f1, dtype: float64\n",
      "END: 1555517098.3156338\n",
      "Duration without docker: 195.68882131576538\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x432 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_params = {**hyperOpt_goodParams, **{'num_epochs': 5, 'language': 'en', 'batch_size': 12, 'embedding_type': 'glove'}}\n",
    "description = '[DOCKER_OFF] This experiment is run to measure the performance decrease when using docker vs. running it locally.'\n",
    "e = Experiment('DockerTest', description, default_params, test_params, dsl, runs=1)\n",
    "\n",
    "start = time.time()\n",
    "print(f'START: {start}')\n",
    "e.run()\n",
    "end = time.time()\n",
    "print(f'END: {end}')\n",
    "duration = end - start\n",
    "print(f'Duration without docker: {duration}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(f'Duration without docker: {duration}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_params = {**hyperOpt_goodParams, **{'num_epochs': 5, 'language': 'en', 'batch_size': 12}}\n",
    "description = '[DOCKER_ON] This experiment is run to measure the performance decrease when using docker vs. running it locally.'\n",
    "e = Experiment('DockerTest', description, default_params, test_params, dsl, runs=1)\n",
    "\n",
    "start = time.time()\n",
    "print(f'START: {start}')\n",
    "e.run()\n",
    "end = time.time()\n",
    "print(f'END: {end}')\n",
    "duration = end - start\n",
    "print(f'Duration with docker: {duration}')"
   ]
  }
 ],
 "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}