notebook/attention_visualization.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "04/04/2019 17:23:39 - INFO - pytorch_pretrained_bert.tokenization -   loading vocabulary file https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-uncased-vocab.txt from cache at /Users/KexinHuang/.pytorch_pretrained_bert/26bc1ad6c0ac742e9b52263248f6d0f00068293b33709fae12320c0e35ccfbbb.542ce4285a40d23a559526243235df47c5f75c197f04f37d1a0c124c32c9a084\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "from torch import nn\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "sns.set_style('white')\n",
    "sns.set_context('talk', font_scale = 1)\n",
    "import os\n",
    "\n",
    "import math\n",
    "from pytorch_pretrained_bert import BertTokenizer\n",
    "os.chdir('../')\n",
    "from modeling_readmission import BertForSequenceClassification\n",
    "import modeling_readmission\n",
    "\n",
    "tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "bert_config = modeling_readmission.BertConfig.from_json_file('../model2/bert_config.json')\n",
    "model = BertForSequenceClassification(bert_config)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{\n",
       "  \"attention_probs_dropout_prob\": 0.1,\n",
       "  \"hidden_act\": \"gelu\",\n",
       "  \"hidden_dropout_prob\": 0.1,\n",
       "  \"hidden_size\": 768,\n",
       "  \"initializer_range\": 0.02,\n",
       "  \"intermediate_size\": 3072,\n",
       "  \"max_position_embeddings\": 512,\n",
       "  \"num_attention_heads\": 12,\n",
       "  \"num_hidden_layers\": 12,\n",
       "  \"type_vocab_size\": 2,\n",
       "  \"vocab_size\": 30522\n",
       "}"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bert_config"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "12"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bert_config.num_attention_heads"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "dicts=model.load_state_dict(torch.load('../model2/pytorch_model.bin',map_location='cpu'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "BertSelfAttention(\n",
       "  (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "  (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "  (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "  (dropout): Dropout(p=0.1)\n",
       ")"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.bert.encoder.layer[0].attention.self"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def transpose_for_scores(config, x):\n",
    "    new_x_shape = x.size()[:-1] + (config.num_attention_heads, int(config.hidden_size / config.num_attention_heads))\n",
    "    x = x.view(*new_x_shape)\n",
    "    return x.permute(0, 2, 1, 3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def get_attention_scores(model,i,text):\n",
    "    tokenized=tokenizer.tokenize(text)\n",
    "\n",
    "    indexed_tokens=tokenizer.convert_tokens_to_ids(tokenized)\n",
    "\n",
    "    segment_ids=[0]*len(indexed_tokens)\n",
    "    t_tensor=torch.tensor([indexed_tokens])\n",
    "    s_ids=torch.tensor([segment_ids])\n",
    "\n",
    "    outputs_query= []\n",
    "    outputs_key= []\n",
    "\n",
    "    def hook_query(module, input, output):\n",
    "        #print ('in query')\n",
    "        outputs_query.append(output)\n",
    "\n",
    "    def hook_key(module, input, output):\n",
    "        #print ('in key')\n",
    "        outputs_key.append(output)\n",
    "\n",
    "    model.bert.encoder.layer[i].attention.self.query.register_forward_hook(hook_query)\n",
    "    model.bert.encoder.layer[i].attention.self.key.register_forward_hook(hook_key)\n",
    "    l=model(t_tensor,s_ids)\n",
    "    \n",
    "    query_layer = transpose_for_scores(bert_config,outputs_query[0])\n",
    "    key_layer = transpose_for_scores(bert_config,outputs_key[0])\n",
    "    \n",
    "    attention_scores = torch.matmul(query_layer, key_layer.transpose(-1, -2))\n",
    "    attention_scores = attention_scores / math.sqrt(int(bert_config.hidden_size / bert_config.num_attention_heads))\n",
    "    attention_probs = nn.Softmax(dim=-1)(attention_scores)\n",
    "    \n",
    "    return attention_probs,tokenized"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "probability:\n",
      "tensor([[[[0.0000, 0.0767, 0.0513,  ..., 0.0344, 0.0000, 0.0383],\n",
      "          [0.0672, 0.0427, 0.0928,  ..., 0.0306, 0.0000, 0.0079],\n",
      "          [0.0660, 0.0755, 0.0487,  ..., 0.0696, 0.0142, 0.0000],\n",
      "          ...,\n",
      "          [0.0101, 0.0216, 0.0000,  ..., 0.0483, 0.0359, 0.0087],\n",
      "          [0.0242, 0.0332, 0.0665,  ..., 0.0753, 0.0136, 0.0125],\n",
      "          [0.0000, 0.0325, 0.0939,  ..., 0.0441, 0.0127, 0.0194]],\n",
      "\n",
      "         [[0.1643, 0.0146, 0.0190,  ..., 0.0131, 0.0261, 0.0067],\n",
      "          [0.0000, 0.0110, 0.0000,  ..., 0.0534, 0.0000, 0.0043],\n",
      "          [0.0000, 0.0240, 0.0129,  ..., 0.0216, 0.0000, 0.0016],\n",
      "          ...,\n",
      "          [0.0127, 0.0000, 0.0232,  ..., 0.0140, 0.1423, 0.0263],\n",
      "          [0.1058, 0.0705, 0.0388,  ..., 0.0658, 0.0103, 0.0297],\n",
      "          [0.0000, 0.1563, 0.0338,  ..., 0.0367, 0.0308, 0.1529]],\n",
      "\n",
      "         [[0.6386, 0.0544, 0.0195,  ..., 0.0000, 0.0123, 0.0000],\n",
      "          [0.8415, 0.0226, 0.0185,  ..., 0.0087, 0.0017, 0.0019],\n",
      "          [0.0000, 0.5207, 0.0266,  ..., 0.0003, 0.0002, 0.0042],\n",
      "          ...,\n",
      "          [0.3249, 0.0029, 0.0008,  ..., 0.0097, 0.0575, 0.0317],\n",
      "          [0.0065, 0.0031, 0.0107,  ..., 0.9827, 0.0004, 0.0070],\n",
      "          [0.0000, 0.0088, 0.0020,  ..., 0.1702, 0.0000, 0.0669]],\n",
      "\n",
      "         ...,\n",
      "\n",
      "         [[0.0662, 0.1513, 0.0000,  ..., 0.0359, 0.0162, 0.0161],\n",
      "          [0.2181, 0.0603, 0.0391,  ..., 0.0190, 0.0413, 0.0168],\n",
      "          [0.0709, 0.0552, 0.1498,  ..., 0.0000, 0.0481, 0.0406],\n",
      "          ...,\n",
      "          [0.1258, 0.0131, 0.0055,  ..., 0.0000, 0.1398, 0.0152],\n",
      "          [0.1353, 0.0068, 0.0098,  ..., 0.0885, 0.3124, 0.0104],\n",
      "          [0.2673, 0.1270, 0.0259,  ..., 0.0256, 0.0203, 0.0137]],\n",
      "\n",
      "         [[0.3318, 0.0200, 0.0397,  ..., 0.0228, 0.0194, 0.0270],\n",
      "          [0.0003, 0.0013, 1.0595,  ..., 0.0000, 0.0002, 0.0006],\n",
      "          [0.0431, 0.0153, 0.0238,  ..., 0.0000, 0.0002, 0.0005],\n",
      "          ...,\n",
      "          [0.0003, 0.0002, 0.0000,  ..., 0.0011, 1.0794, 0.0149],\n",
      "          [0.0174, 0.0000, 0.0001,  ..., 0.0000, 0.0374, 0.9387],\n",
      "          [0.0092, 0.0017, 0.0015,  ..., 0.0481, 0.0000, 0.2650]],\n",
      "\n",
      "         [[0.9432, 0.0047, 0.0125,  ..., 0.0015, 0.0011, 0.0000],\n",
      "          [0.0000, 0.0345, 0.1927,  ..., 0.0047, 0.0000, 0.0125],\n",
      "          [0.1552, 0.1041, 0.0750,  ..., 0.0116, 0.0108, 0.0037],\n",
      "          ...,\n",
      "          [0.0321, 0.0302, 0.0204,  ..., 0.0513, 0.0514, 0.0169],\n",
      "          [0.0085, 0.0045, 0.0098,  ..., 0.0000, 0.0218, 0.0052],\n",
      "          [0.3877, 0.0169, 0.0379,  ..., 0.0375, 0.0404, 0.0019]]]],\n",
      "       grad_fn=<DropoutBackward>)\n",
      "probability:\n",
      "tensor([[[[0.0905, 0.0097, 0.0203,  ..., 0.0226, 0.0000, 0.1059],\n",
      "          [0.0342, 0.1843, 0.1095,  ..., 0.0000, 0.0268, 0.0227],\n",
      "          [0.1375, 0.0191, 0.0034,  ..., 0.0069, 0.0140, 0.3062],\n",
      "          ...,\n",
      "          [0.0832, 0.0198, 0.0000,  ..., 0.0064, 0.0407, 0.0819],\n",
      "          [0.0258, 0.0113, 0.0058,  ..., 0.0112, 0.0261, 0.1165],\n",
      "          [0.0961, 0.0084, 0.0188,  ..., 0.0666, 0.0518, 0.1026]],\n",
      "\n",
      "         [[0.6067, 0.0409, 0.0147,  ..., 0.0132, 0.0160, 0.0411],\n",
      "          [0.1642, 0.0237, 0.5990,  ..., 0.0168, 0.0025, 0.0000],\n",
      "          [0.0000, 0.0416, 0.0268,  ..., 0.0028, 0.0167, 0.0132],\n",
      "          ...,\n",
      "          [0.0000, 0.0015, 0.0001,  ..., 0.0033, 1.0124, 0.0177],\n",
      "          [0.8141, 0.0009, 0.0002,  ..., 0.0149, 0.0179, 0.2004],\n",
      "          [0.5687, 0.0024, 0.0019,  ..., 0.0181, 0.0229, 0.1989]],\n",
      "\n",
      "         [[0.5735, 0.0254, 0.0043,  ..., 0.0120, 0.0158, 0.0000],\n",
      "          [0.2796, 0.0427, 0.0155,  ..., 0.0181, 0.0318, 0.1045],\n",
      "          [0.5546, 0.0059, 0.0009,  ..., 0.0134, 0.0000, 0.0436],\n",
      "          ...,\n",
      "          [0.5320, 0.0087, 0.0031,  ..., 0.0201, 0.0332, 0.0000],\n",
      "          [0.1571, 0.0359, 0.0106,  ..., 0.0195, 0.0153, 0.0479],\n",
      "          [0.0000, 0.0000, 0.0237,  ..., 0.0194, 0.0149, 0.0815]],\n",
      "\n",
      "         ...,\n",
      "\n",
      "         [[0.2194, 0.0564, 0.0371,  ..., 0.0271, 0.0295, 0.0456],\n",
      "          [0.0981, 0.0324, 0.0421,  ..., 0.0237, 0.0140, 0.0761],\n",
      "          [0.1054, 0.1002, 0.0033,  ..., 0.0309, 0.0087, 0.0462],\n",
      "          ...,\n",
      "          [0.2346, 0.0409, 0.0375,  ..., 0.0028, 0.0052, 0.0554],\n",
      "          [0.4464, 0.0267, 0.0000,  ..., 0.0140, 0.0027, 0.0467],\n",
      "          [0.0708, 0.0557, 0.0288,  ..., 0.0536, 0.0224, 0.0509]],\n",
      "\n",
      "         [[0.6948, 0.0000, 0.0094,  ..., 0.0155, 0.0000, 0.0391],\n",
      "          [0.5262, 0.0396, 0.0317,  ..., 0.0044, 0.0035, 0.0110],\n",
      "          [0.0000, 0.1176, 0.0327,  ..., 0.0084, 0.0076, 0.0049],\n",
      "          ...,\n",
      "          [0.2135, 0.0053, 0.0059,  ..., 0.0105, 0.2016, 0.0000],\n",
      "          [0.0132, 0.0045, 0.0060,  ..., 0.1610, 0.0057, 0.0507],\n",
      "          [0.0412, 0.0175, 0.0086,  ..., 0.1039, 0.0218, 0.0537]],\n",
      "\n",
      "         [[0.4985, 0.1132, 0.0185,  ..., 0.0221, 0.0083, 0.0311],\n",
      "          [0.0000, 0.5583, 0.0141,  ..., 0.0017, 0.0029, 0.0000],\n",
      "          [0.0435, 0.0080, 0.6517,  ..., 0.0095, 0.0222, 0.0148],\n",
      "          ...,\n",
      "          [0.1149, 0.0000, 0.0064,  ..., 0.7835, 0.0316, 0.0162],\n",
      "          [0.0164, 0.0012, 0.0128,  ..., 0.0435, 0.9120, 0.0017],\n",
      "          [0.2179, 0.1058, 0.0331,  ..., 0.0533, 0.0124, 0.0095]]]],\n",
      "       grad_fn=<DropoutBackward>)\n",
      "probability:\n",
      "tensor([[[[1.0482, 0.0009, 0.0012,  ..., 0.0004, 0.0007, 0.0119],\n",
      "          [0.0002, 0.0006, 0.0000,  ..., 0.0001, 0.0000, 0.0000],\n",
      "          [0.0006, 0.0000, 0.0000,  ..., 0.0000, 0.0011, 0.0000],\n",
      "          ...,\n",
      "          [0.0000, 0.0000, 0.0000,  ..., 0.0000, 0.0000, 0.0000],\n",
      "          [0.0000, 0.0000, 0.0000,  ..., 0.0000, 0.0000, 1.1098],\n",
      "          [0.0433, 0.0000, 0.0000,  ..., 0.0000, 0.0003, 0.3916]],\n",
      "\n",
      "         [[0.7487, 0.0000, 0.0074,  ..., 0.0110, 0.0061, 0.0625],\n",
      "          [0.9483, 0.0044, 0.0129,  ..., 0.0000, 0.0018, 0.0492],\n",
      "          [0.7729, 0.2671, 0.0012,  ..., 0.0000, 0.0001, 0.0134],\n",
      "          ...,\n",
      "          [0.5276, 0.0045, 0.0003,  ..., 0.0028, 0.0234, 0.1621],\n",
      "          [0.1186, 0.0014, 0.0032,  ..., 0.0000, 0.0060, 0.3495],\n",
      "          [0.6048, 0.0053, 0.0016,  ..., 0.0433, 0.0308, 0.2016]],\n",
      "\n",
      "         [[0.6435, 0.0376, 0.0140,  ..., 0.0000, 0.0249, 0.0000],\n",
      "          [0.7238, 0.0174, 0.0092,  ..., 0.0145, 0.0087, 0.0798],\n",
      "          [0.9381, 0.0000, 0.0000,  ..., 0.0180, 0.0054, 0.0435],\n",
      "          ...,\n",
      "          [0.3076, 0.0126, 0.0799,  ..., 0.0009, 0.0047, 0.0000],\n",
      "          [0.2176, 0.0247, 0.0288,  ..., 0.0268, 0.0034, 0.0958],\n",
      "          [0.5397, 0.0378, 0.0126,  ..., 0.0417, 0.0296, 0.0000]],\n",
      "\n",
      "         ...,\n",
      "\n",
      "         [[0.0000, 0.0176, 0.0146,  ..., 0.0097, 0.0045, 0.0631],\n",
      "          [0.0001, 0.0003, 1.0957,  ..., 0.0002, 0.0000, 0.0000],\n",
      "          [0.0004, 0.0000, 0.0018,  ..., 0.0000, 0.0001, 0.0000],\n",
      "          ...,\n",
      "          [0.0000, 0.0000, 0.0000,  ..., 0.0000, 1.1111, 0.0000],\n",
      "          [0.0000, 0.0000, 0.0000,  ..., 0.0000, 0.0000, 1.1107],\n",
      "          [0.0000, 0.0002, 0.0000,  ..., 0.0024, 0.0003, 0.4147]],\n",
      "\n",
      "         [[0.0000, 0.0428, 0.0088,  ..., 0.0177, 0.0107, 0.0424],\n",
      "          [0.5446, 0.0807, 0.0129,  ..., 0.0497, 0.0000, 0.0390],\n",
      "          [0.0000, 0.0091, 0.0036,  ..., 0.0011, 0.0007, 0.0190],\n",
      "          ...,\n",
      "          [0.5514, 0.0272, 0.0002,  ..., 0.0875, 0.0907, 0.0267],\n",
      "          [0.2039, 0.0198, 0.0019,  ..., 0.2138, 0.0175, 0.1399],\n",
      "          [0.8517, 0.0331, 0.0089,  ..., 0.0115, 0.0098, 0.0238]],\n",
      "\n",
      "         [[0.0000, 0.0000, 0.0017,  ..., 0.0055, 0.0067, 0.0384],\n",
      "          [0.0000, 0.0269, 0.1112,  ..., 0.0142, 0.0575, 0.0984],\n",
      "          [0.3194, 0.0416, 0.0120,  ..., 0.0061, 0.0376, 0.0000],\n",
      "          ...,\n",
      "          [0.7765, 0.0018, 0.0009,  ..., 0.0085, 0.0536, 0.0345],\n",
      "          [0.1603, 0.0207, 0.0277,  ..., 0.0643, 0.0805, 0.1532],\n",
      "          [0.4392, 0.0200, 0.0061,  ..., 0.0133, 0.0289, 0.1523]]]],\n",
      "       grad_fn=<DropoutBackward>)\n",
      "probability:\n",
      "tensor([[[[0.8162, 0.0124, 0.0036,  ..., 0.0005, 0.0007, 0.0000],\n",
      "          [0.2255, 0.8182, 0.0022,  ..., 0.0001, 0.0001, 0.0175],\n",
      "          [0.6006, 0.0028, 0.2336,  ..., 0.0000, 0.0001, 0.0000],\n",
      "          ...,\n",
      "          [0.7139, 0.0002, 0.0000,  ..., 0.3101, 0.0001, 0.0225],\n",
      "          [0.8932, 0.0002, 0.0020,  ..., 0.0001, 0.0641, 0.0200],\n",
      "          [0.8720, 0.0012, 0.0012,  ..., 0.0001, 0.0000, 0.1764]],\n",
      "\n",
      "         [[0.3716, 0.0642, 0.0197,  ..., 0.1027, 0.0163, 0.1009],\n",
      "          [0.1719, 0.1140, 0.0950,  ..., 0.0297, 0.0213, 0.0327],\n",
      "          [0.4687, 0.1093, 0.0181,  ..., 0.0264, 0.0207, 0.0132],\n",
      "          ...,\n",
      "          [0.2247, 0.0253, 0.0000,  ..., 0.0067, 0.0366, 0.0770],\n",
      "          [0.2251, 0.0243, 0.0434,  ..., 0.0274, 0.0149, 0.0398],\n",
      "          [0.3475, 0.0381, 0.0248,  ..., 0.0562, 0.0227, 0.1028]],\n",
      "\n",
      "         [[0.0240, 0.1059, 0.0665,  ..., 0.0647, 0.0431, 0.0108],\n",
      "          [0.3604, 0.0460, 0.0111,  ..., 0.0119, 0.0039, 0.0000],\n",
      "          [0.5243, 0.0189, 0.0117,  ..., 0.0104, 0.0186, 0.0585],\n",
      "          ...,\n",
      "          [0.0926, 0.0000, 0.0000,  ..., 0.0120, 0.0277, 0.0151],\n",
      "          [0.4661, 0.0039, 0.0138,  ..., 0.0063, 0.0042, 0.0333],\n",
      "          [0.4701, 0.0029, 0.0095,  ..., 0.0241, 0.0085, 0.2770]],\n",
      "\n",
      "         ...,\n",
      "\n",
      "         [[0.4178, 0.0167, 0.0223,  ..., 0.0455, 0.0369, 0.0421],\n",
      "          [0.0712, 0.0405, 0.7395,  ..., 0.0069, 0.0000, 0.0016],\n",
      "          [0.0471, 0.0042, 0.0184,  ..., 0.0012, 0.0006, 0.0002],\n",
      "          ...,\n",
      "          [0.0679, 0.0001, 0.0000,  ..., 0.0006, 1.0256, 0.0000],\n",
      "          [0.8553, 0.0001, 0.0000,  ..., 0.0008, 0.0171, 0.1334],\n",
      "          [0.0363, 0.0053, 0.0008,  ..., 0.0000, 0.3173, 0.0000]],\n",
      "\n",
      "         [[0.0000, 0.0163, 0.0122,  ..., 0.0405, 0.0000, 0.0249],\n",
      "          [0.1272, 0.0522, 0.1160,  ..., 0.0092, 0.0414, 0.0099],\n",
      "          [0.3639, 0.0076, 0.0079,  ..., 0.0007, 0.0284, 0.0091],\n",
      "          ...,\n",
      "          [0.9851, 0.0038, 0.0002,  ..., 0.0097, 0.0092, 0.0222],\n",
      "          [0.1393, 0.0000, 0.0001,  ..., 0.6759, 0.0042, 0.0587],\n",
      "          [0.1669, 0.0358, 0.0358,  ..., 0.0777, 0.0687, 0.0354]],\n",
      "\n",
      "         [[0.4081, 0.0256, 0.0731,  ..., 0.0000, 0.0118, 0.0502],\n",
      "          [0.6976, 0.0360, 0.0779,  ..., 0.0012, 0.0006, 0.0195],\n",
      "          [0.5135, 0.1253, 0.1361,  ..., 0.0009, 0.0003, 0.0050],\n",
      "          ...,\n",
      "          [0.0616, 0.0016, 0.0034,  ..., 0.0000, 0.0140, 0.1236],\n",
      "          [0.0047, 0.0015, 0.0015,  ..., 0.0263, 0.0314, 0.0468],\n",
      "          [0.0345, 0.0020, 0.0014,  ..., 0.1522, 0.0341, 0.2902]]]],\n",
      "       grad_fn=<DropoutBackward>)\n",
      "probability:\n",
      "tensor([[[[0.4684, 0.0072, 0.0285,  ..., 0.1312, 0.0710, 0.0056],\n",
      "          [0.0090, 0.0057, 0.0000,  ..., 0.0328, 0.2888, 0.0075],\n",
      "          [0.0150, 0.0046, 0.0068,  ..., 0.0491, 0.1144, 0.0000],\n",
      "          ...,\n",
      "          [0.0531, 0.0058, 0.0062,  ..., 0.1322, 0.4495, 0.0044],\n",
      "          [0.0075, 0.0010, 0.0038,  ..., 0.0995, 0.2320, 0.0007],\n",
      "          [0.1841, 0.0921, 0.0133,  ..., 0.0474, 0.0111, 0.1622]],\n",
      "\n",
      "         [[0.0000, 0.0017, 0.0102,  ..., 0.0157, 0.0275, 0.0054],\n",
      "          [0.0007, 0.0590, 0.0000,  ..., 0.0517, 0.0393, 0.0201],\n",
      "          [0.0071, 0.0460, 0.0309,  ..., 0.0626, 0.0072, 0.0172],\n",
      "          ...,\n",
      "          [0.0224, 0.0738, 0.0385,  ..., 0.0994, 0.0104, 0.0840],\n",
      "          [0.0008, 0.1331, 0.0000,  ..., 0.0455, 0.0014, 0.0322],\n",
      "          [0.0014, 0.0684, 0.0430,  ..., 0.0586, 0.0120, 0.0342]],\n",
      "\n",
      "         [[0.0000, 0.1031, 0.0385,  ..., 0.0326, 0.0639, 0.0632],\n",
      "          [0.4513, 0.0646, 0.0000,  ..., 0.0404, 0.0048, 0.0959],\n",
      "          [0.4943, 0.0553, 0.0000,  ..., 0.1317, 0.0139, 0.0639],\n",
      "          ...,\n",
      "          [0.3712, 0.0424, 0.0418,  ..., 0.0117, 0.0117, 0.1779],\n",
      "          [0.3592, 0.0171, 0.0164,  ..., 0.0000, 0.0008, 0.3145],\n",
      "          [0.2616, 0.1162, 0.0749,  ..., 0.0144, 0.0161, 0.1654]],\n",
      "\n",
      "         ...,\n",
      "\n",
      "         [[0.1168, 0.0306, 0.0813,  ..., 0.0574, 0.0048, 0.0370],\n",
      "          [0.2329, 0.0584, 0.0906,  ..., 0.0216, 0.0068, 0.0695],\n",
      "          [0.0000, 0.0474, 0.0217,  ..., 0.0518, 0.0054, 0.0397],\n",
      "          ...,\n",
      "          [0.1476, 0.0070, 0.0337,  ..., 0.0113, 0.0382, 0.1433],\n",
      "          [0.1955, 0.0072, 0.0130,  ..., 0.1116, 0.0000, 0.0630],\n",
      "          [0.1400, 0.1423, 0.0070,  ..., 0.0816, 0.0447, 0.2549]],\n",
      "\n",
      "         [[0.1924, 0.0544, 0.0642,  ..., 0.0927, 0.0148, 0.1261],\n",
      "          [0.0614, 0.1537, 0.0940,  ..., 0.0334, 0.0315, 0.2255],\n",
      "          [0.1037, 0.1892, 0.1331,  ..., 0.0104, 0.0116, 0.0228],\n",
      "          ...,\n",
      "          [0.0386, 0.0438, 0.0137,  ..., 0.0482, 0.0227, 0.0000],\n",
      "          [0.0608, 0.0000, 0.0018,  ..., 0.1629, 0.0448, 0.2251],\n",
      "          [0.1878, 0.0568, 0.0213,  ..., 0.0894, 0.0162, 0.1434]],\n",
      "\n",
      "         [[0.0952, 0.3174, 0.0305,  ..., 0.0333, 0.0074, 0.0044],\n",
      "          [0.0196, 0.0830, 0.7022,  ..., 0.0254, 0.0004, 0.0134],\n",
      "          [0.1088, 0.0251, 0.0688,  ..., 0.0030, 0.0001, 0.0028],\n",
      "          ...,\n",
      "          [0.0017, 0.0040, 0.0021,  ..., 0.0006, 1.0002, 0.0000],\n",
      "          [0.0133, 0.0000, 0.0002,  ..., 0.0019, 0.0115, 0.9729],\n",
      "          [0.0109, 0.2400, 0.0017,  ..., 0.0373, 0.0671, 0.2170]]]],\n",
      "       grad_fn=<DropoutBackward>)\n",
      "probability:\n",
      "tensor([[[[0.5631, 0.1920, 0.0000,  ..., 0.0156, 0.0255, 0.0120],\n",
      "          [0.0395, 0.0459, 0.2258,  ..., 0.0010, 0.0015, 0.0000],\n",
      "          [0.0150, 0.0162, 0.0732,  ..., 0.0004, 0.0000, 0.0124],\n",
      "          ...,\n",
      "          [0.0860, 0.0126, 0.0092,  ..., 0.0172, 0.0000, 0.4756],\n",
      "          [0.0658, 0.0000, 0.0076,  ..., 0.0301, 0.2383, 0.4845],\n",
      "          [0.0634, 0.2969, 0.0502,  ..., 0.0390, 0.2064, 0.1531]],\n",
      "\n",
      "         [[0.4302, 0.0207, 0.0149,  ..., 0.0048, 0.0624, 0.0000],\n",
      "          [0.0530, 0.8126, 0.0166,  ..., 0.0001, 0.0006, 0.0472],\n",
      "          [0.0000, 0.0000, 0.3367,  ..., 0.0007, 0.0003, 0.0437],\n",
      "          ...,\n",
      "          [0.0000, 0.0029, 0.0004,  ..., 0.0061, 0.0000, 0.0000],\n",
      "          [0.2822, 0.0000, 0.0008,  ..., 0.3103, 0.0407, 0.4191],\n",
      "          [0.0238, 0.0008, 0.0001,  ..., 0.0035, 0.0000, 0.5769]],\n",
      "\n",
      "         [[0.4293, 0.0684, 0.0121,  ..., 0.0035, 0.0297, 0.0313],\n",
      "          [0.0319, 0.1280, 0.0610,  ..., 0.0197, 0.0910, 0.0257],\n",
      "          [0.0849, 0.2273, 0.0813,  ..., 0.0071, 0.0175, 0.0798],\n",
      "          ...,\n",
      "          [0.0965, 0.0817, 0.0130,  ..., 0.0000, 0.0250, 0.0000],\n",
      "          [0.2291, 0.0000, 0.0222,  ..., 0.0059, 0.0236, 0.0102],\n",
      "          [0.1017, 0.2306, 0.0280,  ..., 0.0148, 0.0253, 0.0000]],\n",
      "\n",
      "         ...,\n",
      "\n",
      "         [[0.1422, 0.4421, 0.0303,  ..., 0.0114, 0.0093, 0.0005],\n",
      "          [0.0000, 0.0010, 1.1049,  ..., 0.0000, 0.0000, 0.0000],\n",
      "          [0.0002, 0.0033, 0.0476,  ..., 0.0001, 0.0000, 0.0000],\n",
      "          ...,\n",
      "          [0.0045, 0.0000, 0.0000,  ..., 0.0153, 1.0141, 0.0026],\n",
      "          [0.0228, 0.0002, 0.0000,  ..., 0.0003, 0.0002, 1.0780],\n",
      "          [0.0152, 0.5436, 0.0000,  ..., 0.0000, 0.0738, 0.0638]],\n",
      "\n",
      "         [[0.0000, 0.0000, 0.0141,  ..., 0.0073, 0.1090, 0.1673],\n",
      "          [0.0000, 0.1253, 0.0186,  ..., 0.0009, 0.0044, 0.0741],\n",
      "          [0.0329, 0.7896, 0.1644,  ..., 0.0001, 0.0002, 0.0039],\n",
      "          ...,\n",
      "          [0.0256, 0.0040, 0.0014,  ..., 0.0484, 0.0315, 0.0664],\n",
      "          [0.0091, 0.0006, 0.0004,  ..., 0.0972, 0.0141, 0.0390],\n",
      "          [0.0000, 0.0025, 0.0008,  ..., 0.0825, 0.4940, 0.1881]],\n",
      "\n",
      "         [[0.4977, 0.1859, 0.0000,  ..., 0.0021, 0.0020, 0.1217],\n",
      "          [0.1207, 0.0654, 0.1491,  ..., 0.0011, 0.0006, 0.2123],\n",
      "          [0.1566, 0.0000, 0.0000,  ..., 0.0029, 0.0061, 0.0331],\n",
      "          ...,\n",
      "          [0.0889, 0.0110, 0.0043,  ..., 0.0294, 0.5648, 0.1735],\n",
      "          [0.1228, 0.0000, 0.0304,  ..., 0.0467, 0.1321, 0.3422],\n",
      "          [0.1443, 0.0000, 0.0000,  ..., 0.0032, 0.0000, 0.5098]]]],\n",
      "       grad_fn=<DropoutBackward>)\n",
      "probability:\n",
      "tensor([[[[0.0088, 0.2369, 0.2210,  ..., 0.0000, 0.0000, 0.1247],\n",
      "          [0.0455, 0.0000, 0.0562,  ..., 0.0123, 0.0282, 0.2409],\n",
      "          [0.0195, 0.1302, 0.1763,  ..., 0.0044, 0.0711, 0.0527],\n",
      "          ...,\n",
      "          [0.2047, 0.0425, 0.0080,  ..., 0.0153, 0.0216, 0.0382],\n",
      "          [0.1309, 0.0588, 0.0223,  ..., 0.0000, 0.0256, 0.0200],\n",
      "          [0.0000, 0.2343, 0.0542,  ..., 0.0014, 0.0022, 0.0907]],\n",
      "\n",
      "         [[0.1002, 0.2879, 0.0320,  ..., 0.0119, 0.0181, 0.1652],\n",
      "          [0.0385, 0.0468, 0.0000,  ..., 0.0080, 0.0174, 0.3341],\n",
      "          [0.0624, 0.2994, 0.0990,  ..., 0.0046, 0.0134, 0.0727],\n",
      "          ...,\n",
      "          [0.3708, 0.0914, 0.0368,  ..., 0.0048, 0.0281, 0.0652],\n",
      "          [0.0805, 0.2767, 0.0661,  ..., 0.0050, 0.0461, 0.0448],\n",
      "          [0.0134, 0.5175, 0.0895,  ..., 0.0012, 0.0044, 0.1839]],\n",
      "\n",
      "         [[0.2344, 0.1170, 0.1644,  ..., 0.0017, 0.0038, 0.1507],\n",
      "          [0.0000, 0.3517, 0.1726,  ..., 0.0132, 0.0376, 0.0756],\n",
      "          [0.0199, 0.8219, 0.1365,  ..., 0.0003, 0.0012, 0.0104],\n",
      "          ...,\n",
      "          [0.0333, 0.0363, 0.0024,  ..., 0.0532, 0.0387, 0.1210],\n",
      "          [0.0000, 0.0100, 0.0032,  ..., 0.0256, 0.0423, 0.0000],\n",
      "          [0.0699, 0.0232, 0.0226,  ..., 0.0040, 0.0489, 0.1216]],\n",
      "\n",
      "         ...,\n",
      "\n",
      "         [[0.4595, 0.0228, 0.0224,  ..., 0.0034, 0.0091, 0.0413],\n",
      "          [0.0653, 0.8481, 0.0059,  ..., 0.0001, 0.0000, 0.1653],\n",
      "          [0.0041, 0.9834, 0.1131,  ..., 0.0000, 0.0000, 0.0015],\n",
      "          ...,\n",
      "          [0.0066, 0.0002, 0.0000,  ..., 0.0036, 0.0094, 0.0000],\n",
      "          [0.0012, 0.0000, 0.0001,  ..., 0.0017, 0.0186, 0.0059],\n",
      "          [0.0082, 0.0003, 0.0010,  ..., 0.0010, 0.0394, 0.0000]],\n",
      "\n",
      "         [[0.1066, 0.0144, 0.1606,  ..., 0.0053, 0.0066, 0.0204],\n",
      "          [0.0587, 0.1181, 0.0213,  ..., 0.0008, 0.0000, 0.0120],\n",
      "          [0.1077, 0.3957, 0.1282,  ..., 0.0001, 0.0000, 0.0416],\n",
      "          ...,\n",
      "          [0.2874, 0.0059, 0.0093,  ..., 0.0091, 0.0389, 0.0409],\n",
      "          [0.2038, 0.0012, 0.0022,  ..., 0.0086, 0.0436, 0.0806],\n",
      "          [0.3267, 0.0048, 0.0199,  ..., 0.0092, 0.0000, 0.0752]],\n",
      "\n",
      "         [[0.1292, 0.0356, 0.0015,  ..., 0.0050, 0.0240, 0.6322],\n",
      "          [0.6583, 0.0000, 0.0005,  ..., 0.0000, 0.0003, 0.1812],\n",
      "          [0.0023, 0.9760, 0.1150,  ..., 0.0000, 0.0000, 0.0003],\n",
      "          ...,\n",
      "          [0.0100, 0.0000, 0.0000,  ..., 0.0033, 0.0005, 0.0037],\n",
      "          [0.0015, 0.0000, 0.0000,  ..., 1.0943, 0.0011, 0.0047],\n",
      "          [0.0298, 0.0004, 0.0000,  ..., 0.0007, 0.8137, 0.2534]]]],\n",
      "       grad_fn=<DropoutBackward>)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "probability:\n",
      "tensor([[[[0.0378, 0.2998, 0.0288,  ..., 0.0026, 0.0056, 0.3353],\n",
      "          [0.0236, 0.4910, 0.0357,  ..., 0.0015, 0.0037, 0.3308],\n",
      "          [0.0025, 1.0492, 0.0244,  ..., 0.0000, 0.0001, 0.0118],\n",
      "          ...,\n",
      "          [0.0387, 0.0171, 0.0025,  ..., 0.0129, 0.0044, 0.0344],\n",
      "          [0.0161, 0.0000, 0.0007,  ..., 0.0188, 0.0060, 0.0483],\n",
      "          [0.0150, 0.0543, 0.0088,  ..., 0.0052, 0.0051, 0.1538]],\n",
      "\n",
      "         [[0.2483, 0.0326, 0.0300,  ..., 0.0237, 0.0000, 0.2495],\n",
      "          [0.0088, 0.0652, 0.1884,  ..., 0.0016, 0.0004, 0.0000],\n",
      "          [0.0000, 0.0070, 0.0558,  ..., 0.0001, 0.0000, 0.0017],\n",
      "          ...,\n",
      "          [0.0048, 0.0026, 0.0002,  ..., 0.0000, 0.3849, 0.0000],\n",
      "          [0.0048, 0.0050, 0.0000,  ..., 0.0116, 0.1059, 0.0000],\n",
      "          [0.0214, 0.0359, 0.0049,  ..., 0.0163, 0.0826, 0.5616]],\n",
      "\n",
      "         [[0.2298, 0.0000, 0.0064,  ..., 0.0042, 0.0018, 0.0076],\n",
      "          [0.0003, 0.0403, 0.7779,  ..., 0.0000, 0.0000, 0.0000],\n",
      "          [0.0001, 0.0017, 0.0226,  ..., 0.0000, 0.0000, 0.0000],\n",
      "          ...,\n",
      "          [0.0021, 0.0002, 0.0000,  ..., 0.0291, 0.7403, 0.2264],\n",
      "          [0.0004, 0.0002, 0.0000,  ..., 0.0005, 0.0028, 1.1017],\n",
      "          [0.0220, 0.1443, 0.0010,  ..., 0.0160, 0.0141, 0.6649]],\n",
      "\n",
      "         ...,\n",
      "\n",
      "         [[0.3787, 0.0534, 0.0729,  ..., 0.0014, 0.0103, 0.1107],\n",
      "          [0.1277, 0.6152, 0.0373,  ..., 0.0071, 0.0092, 0.1221],\n",
      "          [0.0035, 0.0000, 0.4823,  ..., 0.0000, 0.0000, 0.0018],\n",
      "          ...,\n",
      "          [0.0031, 0.0000, 0.0003,  ..., 0.0066, 0.0072, 0.0038],\n",
      "          [0.0039, 0.0002, 0.0002,  ..., 0.0170, 0.0000, 0.0207],\n",
      "          [0.0141, 0.0006, 0.0028,  ..., 0.0038, 0.0358, 0.0607]],\n",
      "\n",
      "         [[0.3616, 0.0047, 0.0012,  ..., 0.0692, 0.0350, 0.0000],\n",
      "          [0.0420, 0.0109, 0.0000,  ..., 0.0043, 0.0019, 0.0118],\n",
      "          [0.0116, 0.0000, 0.0603,  ..., 0.0001, 0.0000, 0.0032],\n",
      "          ...,\n",
      "          [0.0010, 0.0006, 0.0000,  ..., 0.0049, 0.6447, 0.4501],\n",
      "          [0.0036, 0.0005, 0.0000,  ..., 0.0115, 0.1086, 0.9799],\n",
      "          [0.0000, 0.0000, 0.0000,  ..., 0.0051, 0.0247, 1.0150]],\n",
      "\n",
      "         [[0.1168, 0.0947, 0.0033,  ..., 0.0000, 0.0084, 0.6992],\n",
      "          [0.0768, 0.4809, 0.0461,  ..., 0.0003, 0.0020, 0.2337],\n",
      "          [0.0023, 0.7867, 0.2996,  ..., 0.0000, 0.0000, 0.0006],\n",
      "          ...,\n",
      "          [0.0032, 0.0018, 0.0015,  ..., 0.0000, 0.0102, 0.1372],\n",
      "          [0.0050, 0.0004, 0.0013,  ..., 0.9451, 0.0000, 0.0916],\n",
      "          [0.0005, 0.0000, 0.0000,  ..., 0.0000, 0.5830, 0.5112]]]],\n",
      "       grad_fn=<DropoutBackward>)\n",
      "probability:\n",
      "tensor([[[[0.0000, 0.0352, 0.0053,  ..., 0.0072, 0.0000, 0.0345],\n",
      "          [0.0420, 0.0327, 0.0934,  ..., 0.0000, 0.0008, 0.0096],\n",
      "          [0.0884, 0.0610, 0.0000,  ..., 0.0010, 0.0001, 0.0056],\n",
      "          ...,\n",
      "          [0.0080, 0.0044, 0.0003,  ..., 0.0000, 0.7763, 0.2057],\n",
      "          [0.0154, 0.0089, 0.0010,  ..., 0.0884, 0.4898, 0.4142],\n",
      "          [0.0000, 0.1069, 0.0136,  ..., 0.0548, 0.0000, 0.3601]],\n",
      "\n",
      "         [[0.0203, 0.1059, 0.0638,  ..., 0.0182, 0.0227, 0.0274],\n",
      "          [0.0623, 0.1483, 0.0230,  ..., 0.0101, 0.0210, 0.0000],\n",
      "          [0.0338, 0.1816, 0.0286,  ..., 0.0544, 0.0748, 0.0375],\n",
      "          ...,\n",
      "          [0.0000, 0.0426, 0.0115,  ..., 0.4529, 0.1270, 0.0787],\n",
      "          [0.0711, 0.0000, 0.0061,  ..., 0.0000, 0.0000, 0.0750],\n",
      "          [0.0314, 0.0228, 0.0130,  ..., 0.0216, 0.0784, 0.1123]],\n",
      "\n",
      "         [[0.0454, 0.5022, 0.0150,  ..., 0.0102, 0.0111, 0.3161],\n",
      "          [0.0010, 0.4000, 0.1727,  ..., 0.0005, 0.0010, 0.0004],\n",
      "          [0.0008, 0.0079, 0.0369,  ..., 0.0011, 0.0002, 0.0015],\n",
      "          ...,\n",
      "          [0.0002, 0.0001, 0.0001,  ..., 0.0014, 0.0000, 0.0019],\n",
      "          [0.0060, 0.0000, 0.0000,  ..., 0.0005, 0.0023, 1.0997],\n",
      "          [0.0223, 0.3569, 0.0010,  ..., 0.0015, 0.0019, 0.6646]],\n",
      "\n",
      "         ...,\n",
      "\n",
      "         [[0.3671, 0.1957, 0.0228,  ..., 0.0136, 0.0057, 0.0614],\n",
      "          [0.0317, 0.1229, 0.0770,  ..., 0.0012, 0.0008, 0.0117],\n",
      "          [0.0274, 0.0474, 0.0140,  ..., 0.0008, 0.0002, 0.0092],\n",
      "          ...,\n",
      "          [0.0061, 0.0099, 0.0000,  ..., 0.0794, 0.2772, 0.4991],\n",
      "          [0.0000, 0.0073, 0.0008,  ..., 0.1588, 0.1023, 0.6382],\n",
      "          [0.0000, 0.1093, 0.0000,  ..., 0.0305, 0.0320, 0.2563]],\n",
      "\n",
      "         [[0.5723, 0.1681, 0.0425,  ..., 0.0281, 0.0000, 0.0083],\n",
      "          [0.0474, 0.5713, 0.0988,  ..., 0.0000, 0.0004, 0.0053],\n",
      "          [0.0313, 0.4961, 0.0512,  ..., 0.0046, 0.0008, 0.0131],\n",
      "          ...,\n",
      "          [0.0380, 0.0333, 0.0068,  ..., 0.2436, 0.0000, 0.0000],\n",
      "          [0.0030, 0.0026, 0.0004,  ..., 0.4287, 0.1220, 0.0576],\n",
      "          [0.0499, 0.0326, 0.0105,  ..., 0.2200, 0.1071, 0.0000]],\n",
      "\n",
      "         [[0.5814, 0.0000, 0.0922,  ..., 0.0000, 0.0007, 0.0049],\n",
      "          [0.0235, 0.4734, 0.0869,  ..., 0.0014, 0.0004, 0.0282],\n",
      "          [0.0000, 0.5840, 0.0000,  ..., 0.0036, 0.0000, 0.0316],\n",
      "          ...,\n",
      "          [0.0096, 0.0203, 0.0053,  ..., 0.0590, 0.0158, 0.0361],\n",
      "          [0.0070, 0.0030, 0.0007,  ..., 0.0583, 0.0071, 0.0164],\n",
      "          [0.0654, 0.1356, 0.0234,  ..., 0.0444, 0.0061, 0.0609]]]],\n",
      "       grad_fn=<DropoutBackward>)\n",
      "probability:\n",
      "tensor([[[[0.0664, 0.4163, 0.0587,  ..., 0.0185, 0.0458, 0.1260],\n",
      "          [0.0108, 0.9031, 0.0734,  ..., 0.0037, 0.0034, 0.0148],\n",
      "          [0.0355, 0.1346, 0.3842,  ..., 0.0192, 0.0178, 0.0134],\n",
      "          ...,\n",
      "          [0.0147, 0.0272, 0.0349,  ..., 0.1366, 0.5801, 0.0381],\n",
      "          [0.0134, 0.0032, 0.0046,  ..., 0.0537, 0.8056, 0.0341],\n",
      "          [0.1481, 0.1452, 0.0134,  ..., 0.0152, 0.0000, 0.2564]],\n",
      "\n",
      "         [[0.0000, 0.0445, 0.0036,  ..., 0.0926, 0.1055, 0.6393],\n",
      "          [0.4521, 0.1941, 0.0375,  ..., 0.0047, 0.0130, 0.0909],\n",
      "          [0.0000, 0.0348, 0.0167,  ..., 0.0012, 0.0108, 0.0413],\n",
      "          ...,\n",
      "          [0.1150, 0.1617, 0.0070,  ..., 0.0032, 0.0000, 0.2073],\n",
      "          [0.2458, 0.0326, 0.0046,  ..., 0.0000, 0.0782, 0.0000],\n",
      "          [0.6482, 0.0236, 0.0014,  ..., 0.0104, 0.0565, 0.1095]],\n",
      "\n",
      "         [[0.0000, 0.7740, 0.0232,  ..., 0.0016, 0.0020, 0.0388],\n",
      "          [0.0034, 1.0205, 0.0000,  ..., 0.0001, 0.0002, 0.0040],\n",
      "          [0.0022, 1.0154, 0.0105,  ..., 0.0002, 0.0003, 0.0024],\n",
      "          ...,\n",
      "          [0.0089, 0.7073, 0.0099,  ..., 0.0055, 0.0282, 0.0000],\n",
      "          [0.0000, 0.5008, 0.0110,  ..., 0.0168, 0.0203, 0.0317],\n",
      "          [0.0057, 0.7408, 0.0118,  ..., 0.0017, 0.0000, 0.0280]],\n",
      "\n",
      "         ...,\n",
      "\n",
      "         [[0.2444, 0.0378, 0.0039,  ..., 0.0043, 0.0471, 0.5803],\n",
      "          [0.0000, 0.1544, 0.0151,  ..., 0.0016, 0.0245, 0.8276],\n",
      "          [0.1052, 0.4035, 0.0867,  ..., 0.0016, 0.0513, 0.0660],\n",
      "          ...,\n",
      "          [0.0942, 0.1665, 0.0192,  ..., 0.0024, 0.0416, 0.0364],\n",
      "          [0.1129, 0.0591, 0.0202,  ..., 0.1638, 0.1455, 0.0569],\n",
      "          [0.0685, 0.0000, 0.0027,  ..., 0.0029, 0.6268, 0.3376]],\n",
      "\n",
      "         [[0.1014, 0.6254, 0.0526,  ..., 0.0015, 0.0212, 0.0859],\n",
      "          [0.0094, 0.4798, 0.0915,  ..., 0.0006, 0.0012, 0.0059],\n",
      "          [0.0766, 0.6977, 0.0240,  ..., 0.0001, 0.0003, 0.0086],\n",
      "          ...,\n",
      "          [0.0000, 0.6148, 0.0227,  ..., 0.0033, 0.0757, 0.1282],\n",
      "          [0.0401, 0.5838, 0.0115,  ..., 0.0000, 0.0216, 0.0000],\n",
      "          [0.0431, 0.7807, 0.0147,  ..., 0.0044, 0.0138, 0.0397]],\n",
      "\n",
      "         [[0.0388, 0.0974, 0.0000,  ..., 0.0315, 0.0461, 0.0267],\n",
      "          [0.0735, 0.3300, 0.0724,  ..., 0.0254, 0.0126, 0.0414],\n",
      "          [0.0846, 0.2205, 0.0270,  ..., 0.0502, 0.2357, 0.0243],\n",
      "          ...,\n",
      "          [0.1085, 0.1426, 0.0588,  ..., 0.0043, 0.0244, 0.0372],\n",
      "          [0.1759, 0.0000, 0.0243,  ..., 0.0154, 0.0257, 0.0851],\n",
      "          [0.1119, 0.0000, 0.0259,  ..., 0.0453, 0.0649, 0.0462]]]],\n",
      "       grad_fn=<DropoutBackward>)\n",
      "probability:\n",
      "tensor([[[[0.0274, 0.5791, 0.0350,  ..., 0.0314, 0.0229, 0.0396],\n",
      "          [0.0495, 0.6712, 0.0291,  ..., 0.0254, 0.0152, 0.0159],\n",
      "          [0.0556, 0.3749, 0.0320,  ..., 0.0518, 0.0403, 0.0208],\n",
      "          ...,\n",
      "          [0.0563, 0.2282, 0.0211,  ..., 0.0719, 0.0000, 0.0316],\n",
      "          [0.1497, 0.1968, 0.0150,  ..., 0.0000, 0.0507, 0.0507],\n",
      "          [0.0393, 0.4001, 0.0305,  ..., 0.0000, 0.0379, 0.0294]],\n",
      "\n",
      "         [[0.1156, 0.0000, 0.0308,  ..., 0.0272, 0.0386, 0.0151],\n",
      "          [0.0348, 0.9058, 0.0000,  ..., 0.0027, 0.0147, 0.0157],\n",
      "          [0.0462, 0.6172, 0.0000,  ..., 0.0081, 0.0342, 0.0194],\n",
      "          ...,\n",
      "          [0.1130, 0.3456, 0.0257,  ..., 0.3512, 0.0000, 0.0535],\n",
      "          [0.0434, 0.1428, 0.0103,  ..., 0.0037, 0.0000, 0.1025],\n",
      "          [0.0796, 0.1565, 0.0000,  ..., 0.0055, 0.2428, 0.2711]],\n",
      "\n",
      "         [[0.0984, 0.4241, 0.2778,  ..., 0.0030, 0.0124, 0.0000],\n",
      "          [0.1060, 0.5935, 0.0951,  ..., 0.0037, 0.0061, 0.0250],\n",
      "          [0.4668, 0.0000, 0.1104,  ..., 0.0033, 0.0046, 0.0173],\n",
      "          ...,\n",
      "          [0.5748, 0.0949, 0.0220,  ..., 0.0180, 0.0659, 0.0763],\n",
      "          [0.6632, 0.0813, 0.0280,  ..., 0.0083, 0.0675, 0.0798],\n",
      "          [0.0000, 0.2659, 0.1126,  ..., 0.0071, 0.0000, 0.0774]],\n",
      "\n",
      "         ...,\n",
      "\n",
      "         [[0.0000, 0.1712, 0.0465,  ..., 0.0019, 0.0000, 0.2213],\n",
      "          [0.0000, 0.9585, 0.0424,  ..., 0.0005, 0.0035, 0.0197],\n",
      "          [0.0330, 0.2649, 0.5100,  ..., 0.0019, 0.0095, 0.0109],\n",
      "          ...,\n",
      "          [0.1022, 0.1961, 0.0302,  ..., 0.3179, 0.0227, 0.0301],\n",
      "          [0.0244, 0.0417, 0.0126,  ..., 0.0004, 0.7792, 0.1620],\n",
      "          [0.0199, 0.0000, 0.0039,  ..., 0.0002, 0.0784, 0.8993]],\n",
      "\n",
      "         [[0.3476, 0.1502, 0.0417,  ..., 0.0229, 0.1429, 0.0566],\n",
      "          [1.0715, 0.0183, 0.0007,  ..., 0.0024, 0.0015, 0.0036],\n",
      "          [1.0853, 0.0026, 0.0014,  ..., 0.0014, 0.0033, 0.0026],\n",
      "          ...,\n",
      "          [1.0012, 0.0005, 0.0000,  ..., 0.0180, 0.0619, 0.0177],\n",
      "          [0.0000, 0.0009, 0.0000,  ..., 0.0000, 0.0795, 0.0000],\n",
      "          [0.9788, 0.0049, 0.0001,  ..., 0.0046, 0.0061, 0.1081]],\n",
      "\n",
      "         [[0.3254, 0.1847, 0.0000,  ..., 0.0000, 0.0527, 0.0481],\n",
      "          [0.0267, 0.2224, 0.0611,  ..., 0.0165, 0.0372, 0.0625],\n",
      "          [0.0976, 0.2496, 0.0483,  ..., 0.0134, 0.0224, 0.0874],\n",
      "          ...,\n",
      "          [0.2514, 0.2080, 0.0096,  ..., 0.0636, 0.0641, 0.2841],\n",
      "          [0.0000, 0.1699, 0.0083,  ..., 0.0264, 0.0458, 0.5491],\n",
      "          [0.0626, 0.2279, 0.0325,  ..., 0.0386, 0.1121, 0.3118]]]],\n",
      "       grad_fn=<DropoutBackward>)\n",
      "probability:\n",
      "tensor([[[[0.8464, 0.0585, 0.0010,  ..., 0.0186, 0.0000, 0.0000],\n",
      "          [0.0000, 0.1035, 0.0030,  ..., 0.0143, 0.0176, 0.0208],\n",
      "          [0.7036, 0.1135, 0.0023,  ..., 0.0132, 0.0229, 0.0229],\n",
      "          ...,\n",
      "          [0.7677, 0.0679, 0.0026,  ..., 0.0117, 0.0206, 0.0200],\n",
      "          [0.6965, 0.0591, 0.0027,  ..., 0.0280, 0.0370, 0.0243],\n",
      "          [0.7037, 0.0778, 0.0024,  ..., 0.0203, 0.0254, 0.0200]],\n",
      "\n",
      "         [[0.0912, 0.0132, 0.0113,  ..., 0.1642, 0.3957, 0.0091],\n",
      "          [0.1456, 0.0309, 0.0000,  ..., 0.1080, 0.2026, 0.0377],\n",
      "          [0.1616, 0.0398, 0.0316,  ..., 0.1004, 0.2234, 0.0385],\n",
      "          ...,\n",
      "          [0.1229, 0.0329, 0.0243,  ..., 0.1167, 0.2776, 0.0347],\n",
      "          [0.1183, 0.0000, 0.0206,  ..., 0.1857, 0.2890, 0.0263],\n",
      "          [0.0979, 0.0000, 0.0000,  ..., 0.0871, 0.3282, 0.0233]],\n",
      "\n",
      "         [[0.0967, 0.0796, 0.0296,  ..., 0.0493, 0.1732, 0.2547],\n",
      "          [0.0882, 0.0000, 0.0000,  ..., 0.0534, 0.1270, 0.2724],\n",
      "          [0.0620, 0.1695, 0.0274,  ..., 0.0473, 0.1216, 0.2393],\n",
      "          ...,\n",
      "          [0.0701, 0.0829, 0.0121,  ..., 0.1425, 0.2081, 0.2275],\n",
      "          [0.0000, 0.0818, 0.0119,  ..., 0.1040, 0.2704, 0.2433],\n",
      "          [0.0868, 0.0975, 0.0266,  ..., 0.0809, 0.1617, 0.2405]],\n",
      "\n",
      "         ...,\n",
      "\n",
      "         [[0.1236, 0.1407, 0.0210,  ..., 0.0193, 0.0000, 0.0358],\n",
      "          [0.1254, 0.1745, 0.0227,  ..., 0.0304, 0.0508, 0.0000],\n",
      "          [0.1871, 0.0000, 0.0155,  ..., 0.0236, 0.0332, 0.0548],\n",
      "          ...,\n",
      "          [0.1616, 0.1231, 0.0141,  ..., 0.0679, 0.0549, 0.0682],\n",
      "          [0.1545, 0.0996, 0.0145,  ..., 0.0000, 0.0959, 0.0583],\n",
      "          [0.1261, 0.0916, 0.0186,  ..., 0.0306, 0.0611, 0.0771]],\n",
      "\n",
      "         [[0.0114, 0.0363, 0.0181,  ..., 0.0313, 0.6484, 0.0612],\n",
      "          [0.0666, 0.0000, 0.0000,  ..., 0.0393, 0.2759, 0.0000],\n",
      "          [0.0626, 0.1196, 0.0219,  ..., 0.0336, 0.2745, 0.1216],\n",
      "          ...,\n",
      "          [0.0580, 0.0712, 0.0129,  ..., 0.0407, 0.0000, 0.0000],\n",
      "          [0.0552, 0.0956, 0.0136,  ..., 0.0370, 0.3779, 0.0876],\n",
      "          [0.0379, 0.1015, 0.0224,  ..., 0.0346, 0.3830, 0.1065]],\n",
      "\n",
      "         [[0.0138, 0.0344, 0.0414,  ..., 0.0152, 0.3976, 0.0000],\n",
      "          [0.0000, 0.0000, 0.0000,  ..., 0.0236, 0.1710, 0.0466],\n",
      "          [0.0431, 0.1363, 0.0349,  ..., 0.0211, 0.1661, 0.0000],\n",
      "          ...,\n",
      "          [0.0278, 0.0728, 0.0133,  ..., 0.0287, 0.2927, 0.0475],\n",
      "          [0.0000, 0.0703, 0.0221,  ..., 0.0000, 0.3337, 0.0666],\n",
      "          [0.0000, 0.0639, 0.0315,  ..., 0.0223, 0.2665, 0.0529]]]],\n",
      "       grad_fn=<DropoutBackward>)\n"
     ]
    }
   ],
   "source": [
    "text=' he has experienced acute on chronic diastolic heart failure in the setting of volume overload due to his sepsis.'\n",
    "x,tokens=get_attention_scores(model,0,text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "24"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "map1=np.asarray(x[0][1].detach().numpy())\n",
    "len(tokens)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x11b36f240>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "f, ax = plt.subplots(figsize=(6,8))\n",
    "sns.heatmap(map1, annot=False, fmt=\"f\", ax=ax, xticklabels = False, yticklabels = False, vmax=0.4, cmap='gray', cbar_kws={'label':'Attention Weight', 'orientation':'horizontal'}, rasterized = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.clf()\n",
    "\n",
    "f=plt.figure(figsize=(10,10))\n",
    "ax = f.add_subplot(1,1,1)\n",
    "i=ax.imshow(map1,interpolation='nearest',cmap='gray')\n",
    "\n",
    "ax.set_yticks(range(len(tokens)))\n",
    "ax.set_yticklabels(tokens)\n",
    "\n",
    "ax.set_xticks(range(len(tokens)))\n",
    "ax.set_xticklabels(tokens,rotation=60)\n",
    "\n",
    "ax.set_xlabel('key')\n",
    "ax.set_ylabel('query')\n",
    "\n",
    "ax.grid(linewidth = 0.8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "text = 'confidential notes from mimic iii'\n",
    "x,tokens=get_attention_scores(model,0,text)\n",
    "map1=np.asarray(x[0][1].detach().numpy())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "f, ax = plt.subplots(figsize=(8,6))\n",
    "sns.heatmap(map1, annot=False, fmt=\"f\", ax=ax, xticklabels = False, yticklabels = False, vmin=0, vmax=0.07, cmap='gray', cbar_kws={'label':'Attention Weight'})\n",
    "\n",
    "ax.set_xlabel('Key Tokens')\n",
    "ax.set_ylabel('Query Tokens')"
   ]
  }
 ],
 "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}