src/notebook.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Generation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.1 Global configuation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "\n",
    "device = \"cuda\"\n",
    "torch.manual_seed(42)\n",
    "torch.use_deterministic_algorithms(True, warn_only=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.2 Load model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from transformers import AutoModelForCausalLM, AutoTokenizer\n",
    "\n",
    "tokenizer = AutoTokenizer.from_pretrained(\"gpt2-medium\")\n",
    "model = AutoModelForCausalLM.from_pretrained(\"gpt2-medium\").to(device)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.3 Configure prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "max_inference_length = 5\n",
    "input_string = \"Super Mario Land is a game that developed by\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.4 Run prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "The attention mask and the pad token id were not set. As a consequence, you may observe unexpected behavior. Please pass your input's `attention_mask` to obtain reliable results.\n",
      "Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "generated full sequence --> ['Super', ' Mario', ' Land', ' is', ' a', ' game', ' that', ' developed', ' by', ' Nintendo', ',', ' and', ' was', ' released']\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/xuan/seq_rationales/.venv/lib/python3.10/site-packages/transformers/models/gpt2/modeling_gpt2.py:184: UserWarning: Deterministic behavior was enabled with either `torch.use_deterministic_algorithms(True)` or `at::Context::setDeterministicAlgorithms(true)`, but this operation is not deterministic because it uses CuBLAS and you have CUDA >= 10.2. To enable deterministic behavior in this case, you must set an environment variable before running your PyTorch application: CUBLAS_WORKSPACE_CONFIG=:4096:8 or CUBLAS_WORKSPACE_CONFIG=:16:8. For more information, go to https://docs.nvidia.com/cuda/cublas/index.html#cublasApi_reproducibility (Triggered internally at ../aten/src/ATen/Context.cpp:159.)\n",
      "  attn_weights = torch.matmul(query, key.transpose(-1, -2))\n",
      "/home/xuan/seq_rationales/.venv/lib/python3.10/site-packages/transformers/models/gpt2/modeling_gpt2.py:219: UserWarning: Deterministic behavior was enabled with either `torch.use_deterministic_algorithms(True)` or `at::Context::setDeterministicAlgorithms(true)`, but this operation is not deterministic because it uses CuBLAS and you have CUDA >= 10.2. To enable deterministic behavior in this case, you must set an environment variable before running your PyTorch application: CUBLAS_WORKSPACE_CONFIG=:4096:8 or CUBLAS_WORKSPACE_CONFIG=:16:8. For more information, go to https://docs.nvidia.com/cuda/cublas/index.html#cublasApi_reproducibility (Triggered internally at ../aten/src/ATen/Context.cpp:159.)\n",
      "  attn_output = torch.matmul(attn_weights, value)\n",
      "/home/xuan/seq_rationales/.venv/lib/python3.10/site-packages/torch/nn/modules/linear.py:114: UserWarning: Deterministic behavior was enabled with either `torch.use_deterministic_algorithms(True)` or `at::Context::setDeterministicAlgorithms(true)`, but this operation is not deterministic because it uses CuBLAS and you have CUDA >= 10.2. To enable deterministic behavior in this case, you must set an environment variable before running your PyTorch application: CUBLAS_WORKSPACE_CONFIG=:4096:8 or CUBLAS_WORKSPACE_CONFIG=:16:8. For more information, go to https://docs.nvidia.com/cuda/cublas/index.html#cublasApi_reproducibility (Triggered internally at ../aten/src/ATen/Context.cpp:159.)\n",
      "  return F.linear(input, self.weight, self.bias)\n"
     ]
    }
   ],
   "source": [
    "input_ids = tokenizer(input_string, return_tensors='pt')['input_ids'][0].to(model.device)\n",
    "generated_ids = model.generate(input_ids=torch.unsqueeze(input_ids, 0), max_length=(input_ids.shape[0] + max_inference_length), do_sample=False)[0]\n",
    "generated_texts = [ tokenizer.decode(token) for token in generated_ids ]\n",
    "print(f'generated full sequence --> {generated_texts}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Rationalization"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.1 Construct rationalizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[POSTagTokenSampler] Loading vocab from tokenizer - 0.00%\n",
      "[POSTagTokenSampler] Loading vocab from tokenizer - 9.95%\n",
      "[POSTagTokenSampler] Loading vocab from tokenizer - 19.90%\n",
      "[POSTagTokenSampler] Loading vocab from tokenizer - 29.85%\n",
      "[POSTagTokenSampler] Loading vocab from tokenizer - 39.80%\n",
      "[POSTagTokenSampler] Loading vocab from tokenizer - 49.74%\n",
      "[POSTagTokenSampler] Loading vocab from tokenizer - 59.69%\n",
      "[POSTagTokenSampler] Loading vocab from tokenizer - 69.64%\n",
      "[POSTagTokenSampler] Loading vocab from tokenizer - 79.59%\n",
      "[POSTagTokenSampler] Loading vocab from tokenizer - 89.54%\n",
      "[POSTagTokenSampler] Loading vocab from tokenizer - 99.49%\n"
     ]
    }
   ],
   "source": [
    "from rationalization.rationalizer.aggregate_rationalizer import AggregateRationalizer\n",
    "from rationalization.rationalizer.importance_score_evaluator.delta_prob import DeltaProbImportanceScoreEvaluator\n",
    "from rationalization.rationalizer.stopping_condition_evaluator.top_k import TopKStoppingConditionEvaluator\n",
    "from rationalization.rationalizer.token_replacement.token_replacer.uniform import UniformTokenReplacer\n",
    "from rationalization.rationalizer.token_replacement.token_sampler.postag import POSTagTokenSampler\n",
    "\n",
    "rational_size = 5\n",
    "rational_size_ratio = None\n",
    "\n",
    "token_sampler = POSTagTokenSampler(tokenizer=tokenizer, device=device)\n",
    "\n",
    "stopping_condition_evaluator = TopKStoppingConditionEvaluator(\n",
    "    model=model, \n",
    "    token_sampler=token_sampler, \n",
    "    top_k=3, \n",
    "    top_n=rational_size, \n",
    "    top_n_ratio=rational_size_ratio, \n",
    "    tokenizer=tokenizer\n",
    ")\n",
    "\n",
    "importance_score_evaluator = DeltaProbImportanceScoreEvaluator(\n",
    "    model=model, \n",
    "    tokenizer=tokenizer, \n",
    "    token_replacer=UniformTokenReplacer(\n",
    "        token_sampler=token_sampler, \n",
    "        ratio=0.3\n",
    "    ),\n",
    "    stopping_condition_evaluator=stopping_condition_evaluator,\n",
    "    max_steps=3000\n",
    ")\n",
    "\n",
    "rationalizer = AggregateRationalizer(\n",
    "    importance_score_evaluator=importance_score_evaluator,\n",
    "    batch_size=8,\n",
    "    overlap_threshold=2,\n",
    "    overlap_strict_pos=True,\n",
    "    top_n=rational_size, \n",
    "    top_n_ratio=rational_size_ratio\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.2 Run rationalization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/xuan/seq_rationales/.venv/lib/python3.10/site-packages/transformers/models/gpt2/modeling_gpt2.py:184: UserWarning: Deterministic behavior was enabled with either `torch.use_deterministic_algorithms(True)` or `at::Context::setDeterministicAlgorithms(true)`, but this operation is not deterministic because it uses CuBLAS and you have CUDA >= 10.2. To enable deterministic behavior in this case, you must set an environment variable before running your PyTorch application: CUBLAS_WORKSPACE_CONFIG=:4096:8 or CUBLAS_WORKSPACE_CONFIG=:16:8. For more information, go to https://docs.nvidia.com/cuda/cublas/index.html#cublasApi_reproducibility (Triggered internally at ../aten/src/ATen/Context.cpp:159.)\n",
      "  attn_weights = torch.matmul(query, key.transpose(-1, -2))\n",
      "/home/xuan/seq_rationales/.venv/lib/python3.10/site-packages/transformers/models/gpt2/modeling_gpt2.py:219: UserWarning: Deterministic behavior was enabled with either `torch.use_deterministic_algorithms(True)` or `at::Context::setDeterministicAlgorithms(true)`, but this operation is not deterministic because it uses CuBLAS and you have CUDA >= 10.2. To enable deterministic behavior in this case, you must set an environment variable before running your PyTorch application: CUBLAS_WORKSPACE_CONFIG=:4096:8 or CUBLAS_WORKSPACE_CONFIG=:16:8. For more information, go to https://docs.nvidia.com/cuda/cublas/index.html#cublasApi_reproducibility (Triggered internally at ../aten/src/ATen/Context.cpp:159.)\n",
      "  attn_output = torch.matmul(attn_weights, value)\n",
      "/home/xuan/seq_rationales/.venv/lib/python3.10/site-packages/torch/nn/modules/linear.py:114: UserWarning: Deterministic behavior was enabled with either `torch.use_deterministic_algorithms(True)` or `at::Context::setDeterministicAlgorithms(true)`, but this operation is not deterministic because it uses CuBLAS and you have CUDA >= 10.2. To enable deterministic behavior in this case, you must set an environment variable before running your PyTorch application: CUBLAS_WORKSPACE_CONFIG=:4096:8 or CUBLAS_WORKSPACE_CONFIG=:16:8. For more information, go to https://docs.nvidia.com/cuda/cublas/index.html#cublasApi_reproducibility (Triggered internally at ../aten/src/ATen/Context.cpp:159.)\n",
      "  return F.linear(input, self.weight, self.bias)\n",
      "/home/xuan/seq_rationales/src/rationalization/rationalizer/aggregate_rationalizer.py:77: UserWarning: _bincount_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at ../aten/src/ATen/Context.cpp:82.)\n",
      "  count_overlap = torch.bincount(pos_top_n.flatten(), minlength=input_ids.shape[1])\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10 / 14\n",
      "Target word     -->  Nintendo\n",
      "Rational pos    --> tensor([0, 1, 2, 3, 4, 5, 6, 7, 8], device='cuda:0')\n",
      "Rational text   --> ['Super', ' Mario', ' Land', ' is', ' a', ' game', ' that', ' developed', ' by']\n",
      "\n",
      "11 / 14\n",
      "Target word     --> ,\n",
      "Rational pos    --> tensor([0, 1, 2, 3, 4, 5, 7, 8, 9], device='cuda:0')\n",
      "Rational text   --> ['Super', ' Mario', ' Land', ' is', ' a', ' game', ' developed', ' by', ' Nintendo']\n",
      "\n",
      "12 / 14\n",
      "Target word     -->  and\n",
      "Rational pos    --> tensor([ 0,  2,  3,  4,  5,  6,  7,  8,  9, 10], device='cuda:0')\n",
      "Rational text   --> ['Super', ' Land', ' is', ' a', ' game', ' that', ' developed', ' by', ' Nintendo', ',']\n",
      "\n",
      "13 / 14\n",
      "Target word     -->  was\n",
      "Rational pos    --> tensor([ 0,  3,  4,  6,  7,  8,  9, 11], device='cuda:0')\n",
      "Rational text   --> ['Super', ' is', ' a', ' that', ' developed', ' by', ' Nintendo', ' and']\n",
      "\n",
      "14 / 14\n",
      "Target word     -->  released\n",
      "Rational pos    --> tensor([ 2,  3,  4,  5,  7,  8,  9, 11, 12], device='cuda:0')\n",
      "Rational text   --> [' Land', ' is', ' a', ' game', ' developed', ' by', ' Nintendo', ' and', ' was']\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# rationalize each generated token\n",
    "\n",
    "importance_scores = []\n",
    "importance_score_map = torch.zeros([generated_ids.shape[0] - input_ids.shape[0], generated_ids.shape[0] - 1], device=device)\n",
    "\n",
    "for target_pos in torch.arange(input_ids.shape[0], generated_ids.shape[0]):\n",
    "    \n",
    "    # extract target\n",
    "    target_id = generated_ids[target_pos]\n",
    "\n",
    "    # rationalization\n",
    "    pos_rational = rationalizer.rationalize(torch.unsqueeze(generated_ids[:target_pos], 0), torch.unsqueeze(target_id, 0))[0]\n",
    "\n",
    "    ids_rational = generated_ids[pos_rational]\n",
    "    text_rational = [ tokenizer.decode([id_rational]) for id_rational in ids_rational ]\n",
    "\n",
    "    importance_score_map[target_pos - input_ids.shape[0], :target_pos] = rationalizer.mean_important_score\n",
    "\n",
    "    print(f'{target_pos + 1} / {generated_ids.shape[0]}')\n",
    "    print(f'Target word     --> {tokenizer.decode(target_id)}', )\n",
    "    print(f\"Rational pos    --> {pos_rational}\")\n",
    "    print(f\"Rational text   --> {text_rational}\")\n",
    "\n",
    "    print()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.3 Visualize rationalization results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='Importance distribution', ylabel='Target'>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 3000x1000 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn\n",
    "seaborn.set(rc={ 'figure.figsize': (30, 10) })\n",
    "s = seaborn.heatmap(\n",
    "    importance_score_map.cpu(), \n",
    "    xticklabels=generated_texts[:-1], \n",
    "    yticklabels=generated_texts[input_ids.shape[0]:], \n",
    "    annot=True, \n",
    "    square=True)\n",
    "s.set_xlabel('Importance distribution')\n",
    "s.set_ylabel('Target')\n",
    "s"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Evaluation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.1 Configure evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "metric_stride = 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.2 Compute metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/xuan/seq_rationales/.venv/lib/python3.10/site-packages/transformers/models/gpt2/modeling_gpt2.py:184: UserWarning: Deterministic behavior was enabled with either `torch.use_deterministic_algorithms(True)` or `at::Context::setDeterministicAlgorithms(true)`, but this operation is not deterministic because it uses CuBLAS and you have CUDA >= 10.2. To enable deterministic behavior in this case, you must set an environment variable before running your PyTorch application: CUBLAS_WORKSPACE_CONFIG=:4096:8 or CUBLAS_WORKSPACE_CONFIG=:16:8. For more information, go to https://docs.nvidia.com/cuda/cublas/index.html#cublasApi_reproducibility (Triggered internally at ../aten/src/ATen/Context.cpp:159.)\n",
      "  attn_weights = torch.matmul(query, key.transpose(-1, -2))\n",
      "/home/xuan/seq_rationales/.venv/lib/python3.10/site-packages/transformers/models/gpt2/modeling_gpt2.py:219: UserWarning: Deterministic behavior was enabled with either `torch.use_deterministic_algorithms(True)` or `at::Context::setDeterministicAlgorithms(true)`, but this operation is not deterministic because it uses CuBLAS and you have CUDA >= 10.2. To enable deterministic behavior in this case, you must set an environment variable before running your PyTorch application: CUBLAS_WORKSPACE_CONFIG=:4096:8 or CUBLAS_WORKSPACE_CONFIG=:16:8. For more information, go to https://docs.nvidia.com/cuda/cublas/index.html#cublasApi_reproducibility (Triggered internally at ../aten/src/ATen/Context.cpp:159.)\n",
      "  attn_output = torch.matmul(attn_weights, value)\n",
      "/home/xuan/seq_rationales/.venv/lib/python3.10/site-packages/torch/nn/modules/linear.py:114: UserWarning: Deterministic behavior was enabled with either `torch.use_deterministic_algorithms(True)` or `at::Context::setDeterministicAlgorithms(true)`, but this operation is not deterministic because it uses CuBLAS and you have CUDA >= 10.2. To enable deterministic behavior in this case, you must set an environment variable before running your PyTorch application: CUBLAS_WORKSPACE_CONFIG=:4096:8 or CUBLAS_WORKSPACE_CONFIG=:16:8. For more information, go to https://docs.nvidia.com/cuda/cublas/index.html#cublasApi_reproducibility (Triggered internally at ../aten/src/ATen/Context.cpp:159.)\n",
      "  return F.linear(input, self.weight, self.bias)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "target_pos: 10, target_token:  Nintendo, Source Soft-NS: 0.0, Source Soft-NC: 0.004445562604814768, Random Soft-NS: 0.0, Random Soft-NC: 0.019402140751481056\n",
      "target_pos: 11, target_token: ,, Source Soft-NS: 0.2553461790084839, Source Soft-NC: 0.01696317456662655, Random Soft-NS: 0.0, Random Soft-NC: 0.006987605709582567\n",
      "target_pos: 12, target_token:  and, Source Soft-NS: 0.0, Source Soft-NC: 0.005914425477385521, Random Soft-NS: 0.003942588809877634, Random Soft-NC: 0.017186634242534637\n",
      "target_pos: 13, target_token:  was, Source Soft-NS: 0.007712581194937229, Source Soft-NC: 0.0076824272982776165, Random Soft-NS: 0.0, Random Soft-NC: 0.005903724115341902\n",
      "target_pos: 14, target_token:  released, Source Soft-NS: 0.02735670655965805, Source Soft-NC: 0.004680009093135595, Random Soft-NS: 0.0, Random Soft-NC: 0.004598025698214769\n",
      "metric_soft_ns: 4.299475193023682, metric_soft_nc: -0.30944153666496277\n"
     ]
    }
   ],
   "source": [
    "from evaluation.evaluator.soft_norm_sufficiency import SoftNormalizedSufficiencyEvaluator\n",
    "from evaluation.evaluator.soft_norm_comprehensiveness import SoftNormalizedComprehensivenessEvaluator\n",
    "soft_norm_suff_evaluator = SoftNormalizedSufficiencyEvaluator(model)\n",
    "soft_norm_comp_evaluator = SoftNormalizedComprehensivenessEvaluator(model)\n",
    "\n",
    "source_soft_ns_all = []\n",
    "source_soft_nc_all = []\n",
    "random_soft_ns_all = []\n",
    "random_soft_nc_all = []\n",
    "target_token_all = []\n",
    "\n",
    "table_details = [ [\"target_pos\", \"target_token\", \"source_soft_ns\", \"source_soft_nc\", \"rand_soft_ns\", \"rand_soft_nc\"] ]\n",
    "\n",
    "for target_pos in torch.arange(input_ids.shape[0], generated_ids.shape[0], metric_stride):\n",
    "\n",
    "    target_token = tokenizer.decode(generated_ids[target_pos])\n",
    "    target_token_all.append(target_token)\n",
    "\n",
    "    input_ids_step = torch.unsqueeze(generated_ids[:target_pos], 0)\n",
    "    target_id_step = torch.unsqueeze(generated_ids[target_pos], 0)\n",
    "    importance_score_step = torch.unsqueeze(importance_score_map[target_pos - input_ids.shape[0], :target_pos], 0)\n",
    "    random_importance_score_step = torch.softmax(torch.rand(importance_score_step.shape, device=device), dim=-1)\n",
    "\n",
    "    # compute Soft-NS and Soft-NC on source importance score\n",
    "\n",
    "    source_soft_ns_step = soft_norm_suff_evaluator.evaluate(input_ids_step, target_id_step, importance_score_step)\n",
    "    source_soft_ns_all.append(source_soft_ns_step)\n",
    "\n",
    "    source_soft_nc_step = soft_norm_comp_evaluator.evaluate(input_ids_step, target_id_step, importance_score_step)\n",
    "    source_soft_nc_all.append(source_soft_nc_step)\n",
    "\n",
    "    # compute Soft-NS and Soft-NC on random importance score\n",
    "\n",
    "    random_soft_ns_step = soft_norm_suff_evaluator.evaluate(input_ids_step, target_id_step, random_importance_score_step)\n",
    "    random_soft_ns_all.append(random_soft_ns_step)\n",
    "\n",
    "    random_soft_nc_step = soft_norm_comp_evaluator.evaluate(input_ids_step, target_id_step, random_importance_score_step)\n",
    "    random_soft_nc_all.append(random_soft_nc_step)\n",
    "\n",
    "    table_details.append([\n",
    "        target_pos.item() + 1, target_token, \n",
    "        f\"{source_soft_ns_step.item():.3f}\", f\"{source_soft_nc_step.item():.3f}\", \n",
    "        f\"{random_soft_ns_step.item():.3f}\", f\"{random_soft_nc_step.item():.3f}\", \n",
    "        # metric_soft_ns_step.item(), metric_soft_nc_step.item()\n",
    "        ])\n",
    "    print(f\"target_pos: {target_pos + 1}, target_token: {target_token}, Source Soft-NS: {source_soft_ns_step}, Source Soft-NC: {source_soft_nc_step}, Random Soft-NS: {random_soft_ns_step}, Random Soft-NC: {random_soft_nc_step}\")\n",
    "\n",
    "# compute metrics on Soft-NS and Soft-NC\n",
    "\n",
    "metric_soft_ns = torch.log(torch.sum(torch.tensor(source_soft_ns_all, device=device)) / torch.sum(torch.tensor(random_soft_ns_all, device=device)))\n",
    "metric_soft_nc = torch.log(torch.sum(torch.tensor(source_soft_nc_all, device=device)) / torch.sum(torch.tensor(random_soft_nc_all, device=device)))\n",
    "\n",
    "print(f\"metric_soft_ns: {metric_soft_ns}, metric_soft_nc: {metric_soft_nc}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.3 Show metrics in tables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<tbody>\n",
       "<tr><td>target_pos</td><td>target_token</td><td>source_soft_ns</td><td>source_soft_nc</td><td>rand_soft_ns</td><td>rand_soft_nc</td></tr>\n",
       "<tr><td>10        </td><td>Nintendo    </td><td>0.000         </td><td>0.004         </td><td>0.000       </td><td>0.019       </td></tr>\n",
       "<tr><td>11        </td><td>,           </td><td>0.255         </td><td>0.017         </td><td>0.000       </td><td>0.007       </td></tr>\n",
       "<tr><td>12        </td><td>and         </td><td>0.000         </td><td>0.006         </td><td>0.004       </td><td>0.017       </td></tr>\n",
       "<tr><td>13        </td><td>was         </td><td>0.008         </td><td>0.008         </td><td>0.000       </td><td>0.006       </td></tr>\n",
       "<tr><td>14        </td><td>released    </td><td>0.027         </td><td>0.005         </td><td>0.000       </td><td>0.005       </td></tr>\n",
       "</tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<tbody>\n",
       "<tr><td>target_tokens                 </td><td>metric_soft_ns</td><td>metric_soft_nc</td></tr>\n",
       "<tr><td>Nintendo$,$ and$ was$ released</td><td>4.299         </td><td>-0.309        </td></tr>\n",
       "</tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import tabulate\n",
    "from IPython.display import HTML, display\n",
    "\n",
    "table_details_html = tabulate.tabulate(table_details, tablefmt='html')\n",
    "\n",
    "display(HTML(table_details_html))\n",
    "\n",
    "table_mean = [\n",
    "        [ \"target_tokens\", \"metric_soft_ns\", \"metric_soft_nc\" ],\n",
    "        [ \"$\".join(target_token_all), f\"{metric_soft_ns.item():.3f}\", f\"{metric_soft_nc.item():.3f}\" ]\n",
    "    ]\n",
    "\n",
    "table_mean_html = tabulate.tabulate(table_mean, tablefmt='html')\n",
    "\n",
    "display(HTML(table_mean_html))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.4 Save results to file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "import csv\n",
    "\n",
    "with open('notebook_details.csv', 'w', newline='') as csvfile:\n",
    "    csvWriter = csv.writer(csvfile)\n",
    "    csvWriter.writerows(table_details)\n",
    "\n",
    "with open('notebook_mean.csv', 'w', newline='') as csvfile:\n",
    "    csvWriter = csv.writer(csvfile)\n",
    "    csvWriter.writerows(table_mean)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".venv",
   "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": "3.10.12"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}