t5-pretrained-question-paraphraser.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install torch==1.4.0\n",
    "!pip install transformers==2.9.0\n",
    "!pip install pytorch_lightning==0.7.5\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Run inference with any question as input"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "device  cpu\n",
      "\n",
      "Original Question ::\n",
      "Which course should I take to get started in data science?\n",
      "\n",
      "\n",
      "Paraphrased Questions :: \n",
      "0: What should I learn to become a data scientist?\n",
      "1: How do I get started with data science?\n",
      "2: How would you start a data science career?\n",
      "3: How can I start learning data science?\n",
      "4: How do you get started in data science?\n",
      "5: What's the best course for data science?\n",
      "6: Which course should I start with for data science?\n",
      "7: What courses should I follow to get started in data science?\n",
      "8: What degree should be taken by a data scientist?\n",
      "9: Which course should I follow to become a Data Scientist?\n"
     ]
    }
   ],
   "source": [
    "\n",
    "import torch\n",
    "from transformers import T5ForConditionalGeneration,T5Tokenizer\n",
    "\n",
    "\n",
    "def set_seed(seed):\n",
    "  torch.manual_seed(seed)\n",
    "  if torch.cuda.is_available():\n",
    "    torch.cuda.manual_seed_all(seed)\n",
    "\n",
    "set_seed(42)\n",
    "\n",
    "model = T5ForConditionalGeneration.from_pretrained('ramsrigouthamg/t5_paraphraser')\n",
    "tokenizer = T5Tokenizer.from_pretrained('t5-base')\n",
    "\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "print (\"device \",device)\n",
    "model = model.to(device)\n",
    "\n",
    "sentence = \"Which course should I take to get started in data science?\"\n",
    "# sentence = \"What are the ingredients required to bake a perfect cake?\"\n",
    "# sentence = \"What is the best possible approach to learn aeronautical engineering?\"\n",
    "# sentence = \"Do apples taste better than oranges in general?\"\n",
    "\n",
    "\n",
    "text =  \"paraphrase: \" + sentence + \" </s>\"\n",
    "\n",
    "\n",
    "max_len = 256\n",
    "\n",
    "encoding = tokenizer.encode_plus(text,pad_to_max_length=True, return_tensors=\"pt\")\n",
    "input_ids, attention_masks = encoding[\"input_ids\"].to(device), encoding[\"attention_mask\"].to(device)\n",
    "\n",
    "\n",
    "# set top_k = 50 and set top_p = 0.95 and num_return_sequences = 3\n",
    "beam_outputs = model.generate(\n",
    "    input_ids=input_ids, attention_mask=attention_masks,\n",
    "    do_sample=True,\n",
    "    max_length=256,\n",
    "    top_k=120,\n",
    "    top_p=0.98,\n",
    "    early_stopping=True,\n",
    "    num_return_sequences=10\n",
    ")\n",
    "\n",
    "\n",
    "print (\"\\nOriginal Question ::\")\n",
    "print (sentence)\n",
    "print (\"\\n\")\n",
    "print (\"Paraphrased Questions :: \")\n",
    "final_outputs =[]\n",
    "for beam_output in beam_outputs:\n",
    "    sent = tokenizer.decode(beam_output, skip_special_tokens=True,clean_up_tokenization_spaces=True)\n",
    "    if sent.lower() != sentence.lower() and sent not in final_outputs:\n",
    "        final_outputs.append(sent)\n",
    "\n",
    "for i, final_output in enumerate(final_outputs):\n",
    "    print(\"{}: {}\".format(i, final_output))\n",
    "\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "conda_pytorch_p36",
   "language": "python",
   "name": "conda_pytorch_p36"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}