figment_torch.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "import h5py\n",
    "import yaml\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from src.Model import *\n",
    "from run import *\n",
    "from src.InferSent import InferSent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "type_embeddings_file = 'data/type_embeddings.pickle'\n",
    "sub_words_emb_file = 'data/_subwords_embeddings.h5py'\n",
    "infer_sent_file = 'data/infersent1.pkl'\n",
    "glove_file = 'data/glove.840B.300d.txt'\n",
    "targets_file = 'data/_targets.h5py'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/alex/miniconda3/envs/ml/lib/python3.7/site-packages/ipykernel_launcher.py:2: YAMLLoadWarning: calling yaml.load() without Loader=... is deprecated, as the default Loader is unsafe. Please read https://msg.pyyaml.org/load for full details.\n",
      "  \n"
     ]
    }
   ],
   "source": [
    "targets = h5py.File(targets_file, 'r')\n",
    "type_to_ix = yaml.load(targets['targets'].attrs['type_to_ix'])\n",
    "types = [t.replace('-', ' ').replace('_', ' ').strip() for t in type_to_ix]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 128(/129) words with w2v vectors\n",
      "Vocab size : 128\n"
     ]
    }
   ],
   "source": [
    "params_model = {'bsize': 64, 'word_emb_dim': 300, 'enc_lstm_dim': 2048,\n",
    "                'pool_type': 'max', 'dpout_model': 0.0, 'version': 1}\n",
    "infersent = InferSent(params_model)\n",
    "infersent.load_state_dict(torch.load(infer_sent_file))\n",
    "infersent.set_w2v_path(glove_file)\n",
    "infersent.build_vocab(types, tokenize=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "embeddings = infersent.encode(types, tokenize=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "(tensor([[ 0.0028, -0.0254,  0.0038,  ..., -0.0096, -0.0073, -0.0289]],\n",
       "        grad_fn=<MaxBackward0>),\n",
       " array([[2, 3, 1, ..., 1, 1, 4]]))"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "infersent.visualize(types[15], tokenize=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = FigmentModel(sub_words_emb_file=sub_words_emb_file, sub_words_num_emb=143123, sub_words_emb_dim=200, clr_num_emb=83, clr_emb_dim=10, seed=23455)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<All keys matched successfully>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.load_state_dict(torch.load('output/model.pt'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "InferSent(\n",
       "  (enc_lstm): LSTM(300, 2048, bidirectional=True)\n",
       ")"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "infersent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}