Simplification with ruGPT3.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "E: Could not open lock file /var/lib/dpkg/lock-frontend - open (13: Permission denied)\n",
      "E: Unable to acquire the dpkg frontend lock (/var/lib/dpkg/lock-frontend), are you root?\n",
      "Requirement already satisfied: cpufeature in /home/user/conda/lib/python3.7/site-packages (0.2.0)\n",
      "Requirement already satisfied: triton==0.2.3 in /home/user/conda/lib/python3.7/site-packages (0.2.3)\n",
      "Requirement already satisfied: numpy in /home/user/conda/lib/python3.7/site-packages (from triton==0.2.3) (1.18.5)\n",
      "Requirement already satisfied: torch in /home/user/conda/lib/python3.7/site-packages (from triton==0.2.3) (1.6.0+cu101)\n",
      "Requirement already satisfied: sympy in /home/user/conda/lib/python3.7/site-packages (from triton==0.2.3) (1.8)\n",
      "Requirement already satisfied: mpmath>=0.19 in /home/user/conda/lib/python3.7/site-packages (from sympy->triton==0.2.3) (1.2.1)\n",
      "Requirement already satisfied: future in /home/user/conda/lib/python3.7/site-packages (from torch->triton==0.2.3) (0.18.2)\n",
      "Requirement already satisfied: deepspeed==0.3.7 in /home/user/conda/lib/python3.7/site-packages (0.3.7)\n",
      "Requirement already satisfied: tensorboardX==1.8 in /home/user/conda/lib/python3.7/site-packages (from deepspeed==0.3.7) (1.8)\n",
      "Requirement already satisfied: tqdm in /home/user/conda/lib/python3.7/site-packages (from deepspeed==0.3.7) (4.60.0)\n",
      "Requirement already satisfied: torchvision>=0.4.0 in /home/user/conda/lib/python3.7/site-packages (from deepspeed==0.3.7) (0.7.0+cu101)\n",
      "Requirement already satisfied: ninja in /home/user/conda/lib/python3.7/site-packages (from deepspeed==0.3.7) (1.10.0.post2)\n",
      "Requirement already satisfied: triton>=0.2.2 in /home/user/conda/lib/python3.7/site-packages (from deepspeed==0.3.7) (0.2.3)\n",
      "Requirement already satisfied: torch>=1.2 in /home/user/conda/lib/python3.7/site-packages (from deepspeed==0.3.7) (1.6.0+cu101)\n",
      "Requirement already satisfied: protobuf>=3.2.0 in /home/user/conda/lib/python3.7/site-packages (from tensorboardX==1.8->deepspeed==0.3.7) (3.15.8)\n",
      "Requirement already satisfied: six in /home/user/conda/lib/python3.7/site-packages (from tensorboardX==1.8->deepspeed==0.3.7) (1.15.0)\n",
      "Requirement already satisfied: numpy in /home/user/conda/lib/python3.7/site-packages (from tensorboardX==1.8->deepspeed==0.3.7) (1.18.5)\n",
      "Requirement already satisfied: future in /home/user/conda/lib/python3.7/site-packages (from torch>=1.2->deepspeed==0.3.7) (0.18.2)\n",
      "Requirement already satisfied: pillow>=4.1.1 in /home/user/conda/lib/python3.7/site-packages (from torchvision>=0.4.0->deepspeed==0.3.7) (8.2.0)\n",
      "Requirement already satisfied: sympy in /home/user/conda/lib/python3.7/site-packages (from triton>=0.2.2->deepspeed==0.3.7) (1.8)\n",
      "Requirement already satisfied: mpmath>=0.19 in /home/user/conda/lib/python3.7/site-packages (from sympy->triton>=0.2.2->deepspeed==0.3.7) (1.2.1)\n"
     ]
    }
   ],
   "source": [
    "!apt-get install llvm-9-dev\n",
    "!pip install cpufeature\n",
    "!pip install triton==0.2.3\n",
    "!DS_BUILD_CPU_ADAM=1 DS_BUILD_SPARSE_ATTN=1 pip install deepspeed==0.3.7"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: transformers in /home/user/conda/lib/python3.7/site-packages (4.6.1)\n",
      "Requirement already satisfied: tqdm>=4.27 in /home/user/conda/lib/python3.7/site-packages (from transformers) (4.60.0)\n",
      "Requirement already satisfied: filelock in /home/user/conda/lib/python3.7/site-packages (from transformers) (3.0.12)\n",
      "Requirement already satisfied: regex!=2019.12.17 in /home/user/conda/lib/python3.7/site-packages (from transformers) (2021.4.4)\n",
      "Requirement already satisfied: tokenizers<0.11,>=0.10.1 in /home/user/conda/lib/python3.7/site-packages (from transformers) (0.10.3)\n",
      "Requirement already satisfied: numpy>=1.17 in /home/user/conda/lib/python3.7/site-packages (from transformers) (1.18.5)\n",
      "Requirement already satisfied: importlib-metadata in /home/user/conda/lib/python3.7/site-packages (from transformers) (3.10.0)\n",
      "Requirement already satisfied: sacremoses in /home/user/conda/lib/python3.7/site-packages (from transformers) (0.0.45)\n",
      "Requirement already satisfied: huggingface-hub==0.0.8 in /home/user/conda/lib/python3.7/site-packages (from transformers) (0.0.8)\n",
      "Requirement already satisfied: requests in /home/user/conda/lib/python3.7/site-packages (from transformers) (2.25.1)\n",
      "Requirement already satisfied: packaging in /home/user/conda/lib/python3.7/site-packages (from transformers) (20.9)\n",
      "Requirement already satisfied: zipp>=0.5 in /home/user/conda/lib/python3.7/site-packages (from importlib-metadata->transformers) (3.4.1)\n",
      "Requirement already satisfied: typing-extensions>=3.6.4 in /home/user/conda/lib/python3.7/site-packages (from importlib-metadata->transformers) (3.7.4.3)\n",
      "Requirement already satisfied: pyparsing>=2.0.2 in /home/user/conda/lib/python3.7/site-packages (from packaging->transformers) (2.4.7)\n",
      "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /home/user/conda/lib/python3.7/site-packages (from requests->transformers) (1.26.4)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in /home/user/conda/lib/python3.7/site-packages (from requests->transformers) (2020.12.5)\n",
      "Requirement already satisfied: chardet<5,>=3.0.2 in /home/user/conda/lib/python3.7/site-packages (from requests->transformers) (4.0.0)\n",
      "Requirement already satisfied: idna<3,>=2.5 in /home/user/conda/lib/python3.7/site-packages (from requests->transformers) (2.10)\n",
      "Requirement already satisfied: click in /home/user/conda/lib/python3.7/site-packages (from sacremoses->transformers) (7.1.2)\n",
      "Requirement already satisfied: joblib in /home/user/conda/lib/python3.7/site-packages (from sacremoses->transformers) (1.0.1)\n",
      "Requirement already satisfied: six in /home/user/conda/lib/python3.7/site-packages (from sacremoses->transformers) (1.15.0)\n",
      "Collecting git+https://github.com/RussianNLP/russian_paraphrasers@master\n",
      "  Cloning https://github.com/RussianNLP/russian_paraphrasers (to revision master) to /tmp/pip-req-build-xnjmva5h\n",
      "  Running command git clone -q https://github.com/RussianNLP/russian_paraphrasers /tmp/pip-req-build-xnjmva5h\n",
      "Collecting nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master\n",
      "  Cloning https://github.com/Maluuba/nlg-eval.git (to revision master) to /tmp/pip-install-t4mnirr8/nlg-eval_be7b67cee49d45e1b5cb93af88d8e202\n",
      "  Running command git clone -q https://github.com/Maluuba/nlg-eval.git /tmp/pip-install-t4mnirr8/nlg-eval_be7b67cee49d45e1b5cb93af88d8e202\n",
      "Requirement already satisfied: nltk in /home/user/conda/lib/python3.7/site-packages (from russian-paraphrasers==0.0.3) (3.6.1)\n",
      "Requirement already satisfied: scipy in /home/user/conda/lib/python3.7/site-packages (from russian-paraphrasers==0.0.3) (1.4.1)\n",
      "Requirement already satisfied: transformers>=4.0.1 in /home/user/conda/lib/python3.7/site-packages (from russian-paraphrasers==0.0.3) (4.6.1)\n",
      "Collecting sentence-transformers==0.4.0\n",
      "  Using cached sentence_transformers-0.4.0-py3-none-any.whl\n",
      "Requirement already satisfied: numpy>=1.11.0 in /home/user/conda/lib/python3.7/site-packages (from nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master->russian-paraphrasers==0.0.3) (1.18.5)\n",
      "Requirement already satisfied: psutil>=5.6.2 in /home/user/conda/lib/python3.7/site-packages (from nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master->russian-paraphrasers==0.0.3) (5.8.0)\n",
      "Requirement already satisfied: requests>=2.19 in /home/user/conda/lib/python3.7/site-packages (from nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master->russian-paraphrasers==0.0.3) (2.25.1)\n",
      "Requirement already satisfied: six>=1.11 in /home/user/conda/lib/python3.7/site-packages (from nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master->russian-paraphrasers==0.0.3) (1.15.0)\n",
      "Requirement already satisfied: Cython>=0.28.5 in /home/user/conda/lib/python3.7/site-packages (from nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master->russian-paraphrasers==0.0.3) (0.29.22)\n",
      "Requirement already satisfied: scikit-learn>=0.17 in /home/user/conda/lib/python3.7/site-packages (from nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master->russian-paraphrasers==0.0.3) (0.24.1)\n",
      "Requirement already satisfied: gensim~=3.8.3 in /home/user/conda/lib/python3.7/site-packages (from nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master->russian-paraphrasers==0.0.3) (3.8.3)\n",
      "Requirement already satisfied: Theano>=0.8.1 in /home/user/conda/lib/python3.7/site-packages (from nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master->russian-paraphrasers==0.0.3) (1.0.5)\n",
      "Requirement already satisfied: tqdm>=4.24 in /home/user/conda/lib/python3.7/site-packages (from nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master->russian-paraphrasers==0.0.3) (4.60.0)\n",
      "Requirement already satisfied: xdg in /home/user/conda/lib/python3.7/site-packages (from nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master->russian-paraphrasers==0.0.3) (5.0.2)\n",
      "Requirement already satisfied: torch>=1.6.0 in /home/user/conda/lib/python3.7/site-packages (from sentence-transformers==0.4.0->russian-paraphrasers==0.0.3) (1.6.0+cu101)\n",
      "Requirement already satisfied: sentencepiece in /home/user/conda/lib/python3.7/site-packages (from sentence-transformers==0.4.0->russian-paraphrasers==0.0.3) (0.1.95)\n",
      "Requirement already satisfied: smart-open>=1.8.1 in /home/user/conda/lib/python3.7/site-packages (from gensim~=3.8.3->nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master->russian-paraphrasers==0.0.3) (5.1.0)\n",
      "Requirement already satisfied: regex in /home/user/conda/lib/python3.7/site-packages (from nltk->russian-paraphrasers==0.0.3) (2021.4.4)\n",
      "Requirement already satisfied: joblib in /home/user/conda/lib/python3.7/site-packages (from nltk->russian-paraphrasers==0.0.3) (1.0.1)\n",
      "Requirement already satisfied: click in /home/user/conda/lib/python3.7/site-packages (from nltk->russian-paraphrasers==0.0.3) (7.1.2)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in /home/user/conda/lib/python3.7/site-packages (from requests>=2.19->nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master->russian-paraphrasers==0.0.3) (2020.12.5)\n",
      "Requirement already satisfied: idna<3,>=2.5 in /home/user/conda/lib/python3.7/site-packages (from requests>=2.19->nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master->russian-paraphrasers==0.0.3) (2.10)\n",
      "Requirement already satisfied: chardet<5,>=3.0.2 in /home/user/conda/lib/python3.7/site-packages (from requests>=2.19->nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master->russian-paraphrasers==0.0.3) (4.0.0)\n",
      "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /home/user/conda/lib/python3.7/site-packages (from requests>=2.19->nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master->russian-paraphrasers==0.0.3) (1.26.4)\n",
      "Requirement already satisfied: threadpoolctl>=2.0.0 in /home/user/conda/lib/python3.7/site-packages (from scikit-learn>=0.17->nlg-eval@ git+https://github.com/Maluuba/nlg-eval.git@master->russian-paraphrasers==0.0.3) (2.1.0)\n",
      "Requirement already satisfied: future in /home/user/conda/lib/python3.7/site-packages (from torch>=1.6.0->sentence-transformers==0.4.0->russian-paraphrasers==0.0.3) (0.18.2)\n",
      "Requirement already satisfied: packaging in /home/user/conda/lib/python3.7/site-packages (from transformers>=4.0.1->russian-paraphrasers==0.0.3) (20.9)\n",
      "Requirement already satisfied: sacremoses in /home/user/conda/lib/python3.7/site-packages (from transformers>=4.0.1->russian-paraphrasers==0.0.3) (0.0.45)\n",
      "Requirement already satisfied: importlib-metadata in /home/user/conda/lib/python3.7/site-packages (from transformers>=4.0.1->russian-paraphrasers==0.0.3) (3.10.0)\n",
      "Requirement already satisfied: filelock in /home/user/conda/lib/python3.7/site-packages (from transformers>=4.0.1->russian-paraphrasers==0.0.3) (3.0.12)\n",
      "Requirement already satisfied: huggingface-hub==0.0.8 in /home/user/conda/lib/python3.7/site-packages (from transformers>=4.0.1->russian-paraphrasers==0.0.3) (0.0.8)\n",
      "Requirement already satisfied: tokenizers<0.11,>=0.10.1 in /home/user/conda/lib/python3.7/site-packages (from transformers>=4.0.1->russian-paraphrasers==0.0.3) (0.10.3)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: typing-extensions>=3.6.4 in /home/user/conda/lib/python3.7/site-packages (from importlib-metadata->transformers>=4.0.1->russian-paraphrasers==0.0.3) (3.7.4.3)\n",
      "Requirement already satisfied: zipp>=0.5 in /home/user/conda/lib/python3.7/site-packages (from importlib-metadata->transformers>=4.0.1->russian-paraphrasers==0.0.3) (3.4.1)\n",
      "Requirement already satisfied: pyparsing>=2.0.2 in /home/user/conda/lib/python3.7/site-packages (from packaging->transformers>=4.0.1->russian-paraphrasers==0.0.3) (2.4.7)\n",
      "Installing collected packages: sentence-transformers\n",
      "  Attempting uninstall: sentence-transformers\n",
      "    Found existing installation: sentence-transformers 1.2.0\n",
      "    Uninstalling sentence-transformers-1.2.0:\n",
      "      Successfully uninstalled sentence-transformers-1.2.0\n",
      "Successfully installed sentence-transformers-0.4.0\n"
     ]
    }
   ],
   "source": [
    "!pip install transformers\n",
    "!pip install git+https://github.com/RussianNLP/russian_paraphrasers@master"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "os.environ[\"USE_DEEPSPEED\"] = \"1\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from src.xl_wrapper import RuGPT3XL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "> initializing model parallel with size 1\n",
      "Use alternating sparse & dense attention layers\n",
      "> initializing model parallel cuda seeds on global rank 0, model parallel rank 0, and data parallel rank 0 with model parallel seed: 3952 and data parallel seed: 1234\n"
     ]
    }
   ],
   "source": [
    "gpt = RuGPT3XL.from_pretrained(\n",
    "    \"sberbank-ai/rugpt3xl\",\n",
    "    weights_path=\"./model/mp_rank_00_model_states.pt\",\n",
    "    deepspeed_config_path=\"./src/deepspeed_config/gpt3_xl_sparse_2048.json\",\n",
    "    seq_len=512\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gpt.tokenizer.add_special_tokens({\"bos_token\": \"<s>\"})\n",
    "gpt.tokenizer.add_special_tokens({\"eos_token\": \"</s>\"})\n",
    "gpt.tokenizer.add_special_tokens({\"sep_token\": \"==>\"})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def generate(text, additional_len=150):\n",
    "    return filter_resuls(gpt.generate(\n",
    "        text=text,\n",
    "        max_length=len(gpt.tokenizer.encode(text)) + additional_len,\n",
    "        eos_token_id=gpt.tokenizer.eos_token_id,\n",
    "        do_sample=True,\n",
    "        num_return_sequences=5,\n",
    "        top_k=10,\n",
    "        top_p=0.9\n",
    "    ))\n",
    "\n",
    "def generate_beam(text, additional_len=150):\n",
    "    return filter_resuls(gpt.generate(\n",
    "        text=text,\n",
    "        max_length=len(gpt.tokenizer.encode(text)) + additional_len,\n",
    "        eos_token_id=gpt.tokenizer.eos_token_id,\n",
    "        num_beams=5,\n",
    "        num_return_sequences=1\n",
    "    ))\n",
    "\n",
    "def generate_nosample(text, additional_len=200):\n",
    "    return filter_resuls(gpt.generate(\n",
    "        text=text,\n",
    "        max_length=len(gpt.tokenizer.encode(text)) + additional_len,\n",
    "        eos_token_id=gpt.tokenizer.eos_token_id,\n",
    "        do_sample=False\n",
    "    ))\n",
    "\n",
    "def filter_resuls(nr):\n",
    "    return [x[:x.find(\"<|endoftext|>\")] for x in nr]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "PyTorch version 1.6.0+cu101 available.\n",
      "TensorFlow version 2.3.0 available.\n"
     ]
    }
   ],
   "source": [
    "from russian_paraphrasers import GPTParaphraser\n",
    "import difflib\n",
    "import re\n",
    "\n",
    "paraphraser = GPTParaphraser(model_name=\"gpt3\", range_cand=True, make_eval=False)\n",
    "\n",
    "\n",
    "def check_bad_generations(origin, pred):\n",
    "    bad = False\n",
    "    origin = origin.strip()\n",
    "    pred = pred.strip()\n",
    "    ratio = difflib.SequenceMatcher(None, origin.lower(), pred.lower()).ratio()\n",
    "    if ratio > 0.97:\n",
    "#         print(\"REPEAT!\")\n",
    "#         print(origin, \"===> \", pred.strip())\n",
    "        bad  = True\n",
    "\n",
    "    # Length\n",
    "    if len(origin) - len(pred) > len(origin) - 50:\n",
    "#         print(\"Length: \", origin, \"===> \", pred)\n",
    "        bad = True\n",
    "        \n",
    "    # check separation union\n",
    "    origin_t = re.sub(\"[^а-я ]\", \"\", origin.lower())\n",
    "    or_tokens = set(origin_t.split())\n",
    "\n",
    "    pred_t = re.sub(\"[^а-я ]\", \"\", pred.lower())\n",
    "    pr_tokens = set(pred_t.split())\n",
    "\n",
    "#   Maybe without this part!\n",
    "    sep = (pr_tokens & or_tokens)\n",
    "    if len(sep) < 3:\n",
    "        # print(\"Smth strange:\", pr_tokens, or_tokens)\n",
    "#         print(\"or: \", origin)\n",
    "#         print(\"===> \", pred)\n",
    "        bad = True\n",
    "    return bad\n",
    "\n",
    "def get_paraphraser_var(origin):\n",
    "    results = paraphraser.generate(\n",
    "            origin, n=10, temperature=1, \n",
    "            top_k=10, top_p=0.9, \n",
    "            max_length=200, repetition_penalty=1.5,\n",
    "            threshold=0.8\n",
    "        )\n",
    "    try:\n",
    "        res = results[\"results\"][0][\"best_candidates\"][0]\n",
    "        if res.endswith(\"?\"):\n",
    "            res = results[\"results\"][0][\"best_candidates\"][1]\n",
    "    except IndexError:\n",
    "            res = results[\"results\"][0][\"predictions\"][0]\n",
    "    return res"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|          | 0/1126 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1126\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  0%|          | 1/1126 [00:09<2:58:13,  9.51s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  0%|          | 2/1126 [00:15<2:23:04,  7.64s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  0%|          | 4/1126 [00:21<1:20:53,  4.33s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  0%|          | 5/1126 [00:25<1:18:41,  4.21s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  1%|          | 8/1126 [00:33<50:04,  2.69s/it]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  1%|          | 11/1126 [00:40<39:23,  2.12s/it] Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  1%|          | 12/1126 [00:45<54:09,  2.92s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  1%|▏         | 15/1126 [00:51<37:26,  2.02s/it]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  2%|▏         | 21/1126 [00:59<19:10,  1.04s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  2%|▏         | 22/1126 [01:04<38:33,  2.10s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  2%|▏         | 23/1126 [01:08<50:43,  2.76s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  2%|▏         | 24/1126 [01:12<57:46,  3.15s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  2%|▏         | 25/1126 [01:16<1:03:41,  3.47s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  2%|▏         | 27/1126 [01:22<53:43,  2.93s/it]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  3%|▎         | 29/1126 [01:27<46:39,  2.55s/it]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  3%|▎         | 30/1126 [01:31<53:54,  2.95s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  3%|▎         | 31/1126 [01:34<58:01,  3.18s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  3%|▎         | 35/1126 [01:41<30:11,  1.66s/it]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  3%|▎         | 38/1126 [01:46<27:13,  1.50s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  4%|▎         | 40/1126 [01:52<34:33,  1.91s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  4%|▎         | 41/1126 [01:56<48:36,  2.69s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  4%|▎         | 42/1126 [02:00<54:57,  3.04s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  4%|▍         | 43/1126 [02:04<58:26,  3.24s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  4%|▍         | 45/1126 [02:10<53:19,  2.96s/it]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  4%|▍         | 46/1126 [02:14<1:03:16,  3.51s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  4%|▍         | 47/1126 [02:18<1:04:37,  3.59s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  4%|▍         | 48/1126 [02:23<1:09:55,  3.89s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  4%|▍         | 50/1126 [02:27<52:25,  2.92s/it]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  5%|▍         | 52/1126 [02:32<44:33,  2.49s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  5%|▍         | 53/1126 [02:36<52:30,  2.94s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  6%|▌         | 62/1126 [02:47<14:38,  1.21it/s]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  6%|▌         | 66/1126 [02:54<21:44,  1.23s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  6%|▋         | 71/1126 [03:01<16:00,  1.10it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  7%|▋         | 76/1126 [03:08<16:57,  1.03it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  7%|▋         | 79/1126 [03:13<22:50,  1.31s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  7%|▋         | 83/1126 [03:19<17:48,  1.02s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  7%|▋         | 84/1126 [03:23<32:44,  1.89s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  8%|▊         | 86/1126 [03:28<35:08,  2.03s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  8%|▊         | 87/1126 [03:31<44:08,  2.55s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  8%|▊         | 88/1126 [03:37<59:07,  3.42s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  8%|▊         | 90/1126 [03:43<52:03,  3.02s/it]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  8%|▊         | 91/1126 [03:48<1:01:22,  3.56s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  8%|▊         | 94/1126 [03:54<39:10,  2.28s/it]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  9%|▊         | 97/1126 [04:00<30:51,  1.80s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  9%|▊         | 98/1126 [04:04<43:02,  2.51s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  9%|▉         | 99/1126 [04:08<49:52,  2.91s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  9%|▉         | 100/1126 [04:13<1:00:27,  3.54s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  9%|▉         | 101/1126 [04:18<1:08:53,  4.03s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  9%|▉         | 102/1126 [04:23<1:11:58,  4.22s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  9%|▉         | 104/1126 [04:28<55:36,  3.26s/it]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  9%|▉         | 105/1126 [04:33<1:03:57,  3.76s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "  9%|▉         | 106/1126 [04:39<1:12:54,  4.29s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 10%|▉         | 107/1126 [04:43<1:15:43,  4.46s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 10%|▉         | 108/1126 [04:48<1:18:26,  4.62s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 10%|▉         | 111/1126 [04:55<47:03,  2.78s/it]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 10%|█         | 114/1126 [05:02<36:53,  2.19s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 10%|█         | 117/1126 [05:08<29:04,  1.73s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 10%|█         | 118/1126 [05:13<44:23,  2.64s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 11%|█         | 119/1126 [05:18<55:54,  3.33s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 11%|█         | 123/1126 [05:25<29:09,  1.74s/it]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 12%|█▏        | 130/1126 [05:33<12:37,  1.31it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 12%|█▏        | 133/1126 [05:37<17:16,  1.04s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 12%|█▏        | 134/1126 [05:41<31:38,  1.91s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 12%|█▏        | 135/1126 [05:45<42:01,  2.54s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 12%|█▏        | 136/1126 [05:50<51:05,  3.10s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 12%|█▏        | 137/1126 [05:54<54:43,  3.32s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 12%|█▏        | 138/1126 [05:58<57:34,  3.50s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 12%|█▏        | 140/1126 [06:04<50:04,  3.05s/it]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 13%|█▎        | 141/1126 [06:09<1:01:13,  3.73s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 13%|█▎        | 144/1126 [06:17<42:43,  2.61s/it]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 13%|█▎        | 145/1126 [06:21<48:39,  2.98s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 13%|█▎        | 146/1126 [06:24<52:47,  3.23s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 14%|█▎        | 153/1126 [06:32<14:41,  1.10it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 14%|█▍        | 156/1126 [06:42<29:13,  1.81s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 14%|█▍        | 158/1126 [06:47<33:51,  2.10s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 14%|█▍        | 159/1126 [06:53<48:44,  3.02s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 14%|█▍        | 160/1126 [06:58<58:35,  3.64s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 15%|█▌        | 170/1126 [07:09<11:10,  1.43it/s]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 15%|█▌        | 174/1126 [07:15<15:52,  1.00s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 16%|█▌        | 175/1126 [07:19<30:18,  1.91s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 16%|█▌        | 181/1126 [07:26<14:27,  1.09it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 17%|█▋        | 187/1126 [07:33<12:27,  1.26it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 17%|█▋        | 189/1126 [07:38<21:15,  1.36s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 17%|█▋        | 193/1126 [07:43<16:58,  1.09s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 17%|█▋        | 194/1126 [07:47<31:23,  2.02s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 17%|█▋        | 195/1126 [07:52<41:51,  2.70s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 17%|█▋        | 196/1126 [07:56<49:27,  3.19s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 18%|█▊        | 201/1126 [08:03<20:05,  1.30s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 18%|█▊        | 202/1126 [08:07<32:52,  2.14s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 18%|█▊        | 205/1126 [08:11<22:41,  1.48s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 18%|█▊        | 206/1126 [08:15<34:48,  2.27s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 18%|█▊        | 207/1126 [08:20<43:49,  2.86s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 18%|█▊        | 208/1126 [08:24<49:12,  3.22s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 19%|█▊        | 211/1126 [08:29<30:17,  1.99s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 19%|█▉        | 216/1126 [08:38<21:15,  1.40s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 19%|█▉        | 217/1126 [08:42<36:12,  2.39s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 19%|█▉        | 218/1126 [08:47<48:08,  3.18s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 20%|█▉        | 220/1126 [08:53<41:33,  2.75s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 20%|█▉        | 221/1126 [08:58<52:46,  3.50s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 20%|█▉        | 222/1126 [09:02<54:55,  3.64s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 20%|██        | 226/1126 [09:08<24:30,  1.63s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 20%|██        | 227/1126 [09:12<35:41,  2.38s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 20%|██        | 228/1126 [09:16<42:27,  2.84s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 20%|██        | 230/1126 [09:21<36:31,  2.45s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 21%|██        | 234/1126 [09:27<21:35,  1.45s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 21%|██        | 236/1126 [09:31<25:41,  1.73s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 21%|██        | 237/1126 [09:38<48:31,  3.27s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 21%|██        | 238/1126 [09:43<53:56,  3.65s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 21%|██▏       | 241/1126 [09:48<33:05,  2.24s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 22%|██▏       | 243/1126 [09:53<31:34,  2.15s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 22%|██▏       | 244/1126 [09:57<42:21,  2.88s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 22%|██▏       | 245/1126 [10:01<47:26,  3.23s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 22%|██▏       | 246/1126 [10:05<50:03,  3.41s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 22%|██▏       | 247/1126 [10:10<57:08,  3.90s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 22%|██▏       | 248/1126 [10:16<1:03:19,  4.33s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 22%|██▏       | 250/1126 [10:22<51:08,  3.50s/it]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 22%|██▏       | 251/1126 [10:26<55:20,  3.80s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 22%|██▏       | 252/1126 [10:30<56:43,  3.89s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 22%|██▏       | 253/1126 [10:34<56:52,  3.91s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 23%|██▎       | 254/1126 [10:39<1:00:29,  4.16s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 23%|██▎       | 256/1126 [10:45<47:39,  3.29s/it]  Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 23%|██▎       | 257/1126 [10:50<54:18,  3.75s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 23%|██▎       | 259/1126 [10:55<44:30,  3.08s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 23%|██▎       | 262/1126 [11:01<29:56,  2.08s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 24%|██▎       | 266/1126 [11:07<18:20,  1.28s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 24%|██▍       | 272/1126 [11:14<14:13,  1.00it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 24%|██▍       | 273/1126 [11:18<27:05,  1.91s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 24%|██▍       | 275/1126 [11:23<27:15,  1.92s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 25%|██▍       | 280/1126 [11:29<13:52,  1.02it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 25%|██▍       | 281/1126 [11:33<25:54,  1.84s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 25%|██▌       | 282/1126 [11:37<36:36,  2.60s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 25%|██▌       | 283/1126 [11:41<41:46,  2.97s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 25%|██▌       | 287/1126 [11:46<20:41,  1.48s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 26%|██▌       | 288/1126 [11:50<28:37,  2.05s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 26%|██▌       | 291/1126 [11:55<23:07,  1.66s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 26%|██▋       | 298/1126 [12:05<11:57,  1.15it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 27%|██▋       | 301/1126 [12:11<19:33,  1.42s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 27%|██▋       | 302/1126 [12:13<23:07,  1.68s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 27%|██▋       | 303/1126 [12:17<31:04,  2.27s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 27%|██▋       | 306/1126 [12:22<22:46,  1.67s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 28%|██▊       | 312/1126 [12:29<13:09,  1.03it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 28%|██▊       | 313/1126 [12:34<28:46,  2.12s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 28%|██▊       | 314/1126 [12:40<42:14,  3.12s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 28%|██▊       | 315/1126 [12:44<48:58,  3.62s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 28%|██▊       | 316/1126 [12:48<50:15,  3.72s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 28%|██▊       | 317/1126 [12:53<51:52,  3.85s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 28%|██▊       | 320/1126 [12:58<29:11,  2.17s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 29%|██▊       | 321/1126 [13:02<37:39,  2.81s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 29%|██▊       | 322/1126 [13:06<42:47,  3.19s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 29%|██▊       | 323/1126 [13:10<46:10,  3.45s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 29%|██▉       | 325/1126 [13:16<39:39,  2.97s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 29%|██▉       | 326/1126 [13:20<43:19,  3.25s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 29%|██▉       | 330/1126 [13:27<23:28,  1.77s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 29%|██▉       | 331/1126 [13:31<34:16,  2.59s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 30%|██▉       | 333/1126 [13:37<32:50,  2.49s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 30%|██▉       | 335/1126 [13:41<28:08,  2.13s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 30%|██▉       | 336/1126 [13:45<36:57,  2.81s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 30%|██▉       | 337/1126 [13:49<40:28,  3.08s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 30%|███       | 338/1126 [13:53<44:44,  3.41s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 30%|███       | 339/1126 [13:57<47:27,  3.62s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 30%|███       | 340/1126 [14:02<51:20,  3.92s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 30%|███       | 342/1126 [14:07<38:24,  2.94s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 31%|███       | 345/1126 [14:11<22:34,  1.73s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 31%|███       | 346/1126 [14:15<30:20,  2.33s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 31%|███▏      | 352/1126 [14:22<12:05,  1.07it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 31%|███▏      | 354/1126 [14:27<18:51,  1.47s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 32%|███▏      | 360/1126 [14:35<12:42,  1.00it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 32%|███▏      | 361/1126 [14:39<24:36,  1.93s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 32%|███▏      | 362/1126 [14:43<31:47,  2.50s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 32%|███▏      | 364/1126 [14:47<27:48,  2.19s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 32%|███▏      | 365/1126 [14:51<33:18,  2.63s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 33%|███▎      | 368/1126 [14:57<23:57,  1.90s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 33%|███▎      | 369/1126 [15:01<32:59,  2.61s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 33%|███▎      | 373/1126 [15:07<17:29,  1.39s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 33%|███▎      | 376/1126 [15:12<18:04,  1.45s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 33%|███▎      | 377/1126 [15:17<31:03,  2.49s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 34%|███▎      | 378/1126 [15:22<40:41,  3.26s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 34%|███▎      | 379/1126 [15:27<46:01,  3.70s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 34%|███▎      | 380/1126 [15:32<50:22,  4.05s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 34%|███▍      | 387/1126 [15:41<12:49,  1.04s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 34%|███▍      | 388/1126 [15:46<28:34,  2.32s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 35%|███▍      | 389/1126 [15:51<38:13,  3.11s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 35%|███▍      | 390/1126 [15:56<44:48,  3.65s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 35%|███▌      | 398/1126 [16:07<14:15,  1.18s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 35%|███▌      | 399/1126 [16:12<27:34,  2.28s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 36%|███▌      | 400/1126 [16:17<36:47,  3.04s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 36%|███▌      | 402/1126 [16:21<30:31,  2.53s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 36%|███▌      | 403/1126 [16:26<36:16,  3.01s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 36%|███▌      | 408/1126 [16:32<16:03,  1.34s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 37%|███▋      | 415/1126 [16:40<08:44,  1.36it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 37%|███▋      | 420/1126 [16:48<11:58,  1.02s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 37%|███▋      | 421/1126 [16:51<22:24,  1.91s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 37%|███▋      | 422/1126 [16:57<33:26,  2.85s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 38%|███▊      | 424/1126 [17:01<29:06,  2.49s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 38%|███▊      | 427/1126 [17:06<20:21,  1.75s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 38%|███▊      | 428/1126 [17:10<27:30,  2.36s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 38%|███▊      | 430/1126 [17:15<25:31,  2.20s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 39%|███▊      | 434/1126 [17:21<17:02,  1.48s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 39%|███▊      | 435/1126 [17:26<27:37,  2.40s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 39%|███▉      | 438/1126 [17:32<20:55,  1.83s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 39%|███▉      | 439/1126 [17:36<30:37,  2.68s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 39%|███▉      | 442/1126 [17:41<19:53,  1.75s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 39%|███▉      | 443/1126 [17:45<27:15,  2.40s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 39%|███▉      | 444/1126 [17:49<34:01,  2.99s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 40%|███▉      | 445/1126 [17:53<37:22,  3.29s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 40%|███▉      | 447/1126 [17:58<29:44,  2.63s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 41%|████      | 458/1126 [18:09<10:07,  1.10it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 41%|████      | 459/1126 [18:15<23:50,  2.14s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 41%|████      | 460/1126 [18:19<33:02,  2.98s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 41%|████      | 463/1126 [18:26<23:30,  2.13s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 41%|████      | 464/1126 [18:31<31:44,  2.88s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 42%|████▏     | 468/1126 [18:38<17:44,  1.62s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 42%|████▏     | 471/1126 [18:45<19:04,  1.75s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 42%|████▏     | 473/1126 [18:50<22:34,  2.07s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 42%|████▏     | 474/1126 [18:55<32:43,  3.01s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 42%|████▏     | 478/1126 [19:01<15:57,  1.48s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 43%|████▎     | 484/1126 [19:09<10:19,  1.04it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 43%|████▎     | 485/1126 [19:14<23:43,  2.22s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 43%|████▎     | 489/1126 [19:22<17:46,  1.67s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 44%|████▎     | 490/1126 [19:27<28:31,  2.69s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 44%|████▍     | 493/1126 [19:33<21:17,  2.02s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 44%|████▍     | 494/1126 [19:39<32:38,  3.10s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 44%|████▍     | 500/1126 [19:49<15:04,  1.44s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 44%|████▍     | 501/1126 [19:53<25:32,  2.45s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 45%|████▍     | 503/1126 [20:00<28:13,  2.72s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 45%|████▍     | 504/1126 [20:06<37:03,  3.57s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 45%|████▌     | 509/1126 [20:14<15:35,  1.52s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 45%|████▌     | 510/1126 [20:20<26:36,  2.59s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 46%|████▌     | 514/1126 [20:28<17:33,  1.72s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 46%|████▌     | 515/1126 [20:32<27:02,  2.66s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 46%|████▌     | 516/1126 [20:38<34:39,  3.41s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 46%|████▌     | 518/1126 [20:43<28:14,  2.79s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 46%|████▌     | 519/1126 [20:46<31:15,  3.09s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 46%|████▌     | 520/1126 [20:55<48:06,  4.76s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 46%|████▋     | 522/1126 [21:00<34:41,  3.45s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 46%|████▋     | 523/1126 [21:04<37:12,  3.70s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 47%|████▋     | 524/1126 [21:10<41:24,  4.13s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 47%|████▋     | 525/1126 [21:14<43:15,  4.32s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 47%|████▋     | 526/1126 [21:19<45:17,  4.53s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 47%|████▋     | 528/1126 [21:25<35:06,  3.52s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 47%|████▋     | 530/1126 [21:31<29:26,  2.96s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 47%|████▋     | 533/1126 [21:37<20:14,  2.05s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 48%|████▊     | 535/1126 [21:42<20:48,  2.11s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 48%|████▊     | 539/1126 [21:48<13:34,  1.39s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 48%|████▊     | 546/1126 [21:56<08:18,  1.16it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 49%|████▊     | 547/1126 [22:00<17:17,  1.79s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 49%|████▊     | 548/1126 [22:08<35:05,  3.64s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 49%|████▉     | 549/1126 [22:12<37:22,  3.89s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 49%|████▉     | 550/1126 [22:16<38:15,  3.98s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 49%|████▉     | 551/1126 [22:24<49:17,  5.14s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 49%|████▉     | 552/1126 [22:28<45:22,  4.74s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 49%|████▉     | 553/1126 [22:32<42:21,  4.44s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 49%|████▉     | 554/1126 [22:35<40:21,  4.23s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 50%|████▉     | 558/1126 [22:41<17:21,  1.83s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 50%|█████     | 563/1126 [22:48<10:21,  1.10s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 50%|█████     | 564/1126 [22:52<18:58,  2.03s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 50%|█████     | 568/1126 [22:58<12:44,  1.37s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 51%|█████     | 571/1126 [23:04<12:36,  1.36s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 51%|█████     | 572/1126 [23:08<19:58,  2.16s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 51%|█████     | 575/1126 [23:12<13:40,  1.49s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 51%|█████     | 576/1126 [23:16<20:44,  2.26s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 51%|█████     | 577/1126 [23:20<24:58,  2.73s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 52%|█████▏    | 581/1126 [23:26<13:15,  1.46s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 52%|█████▏    | 583/1126 [23:31<16:41,  1.84s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 52%|█████▏    | 584/1126 [23:34<21:22,  2.37s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 52%|█████▏    | 591/1126 [23:43<09:18,  1.04s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 53%|█████▎    | 592/1126 [23:46<16:17,  1.83s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 53%|█████▎    | 601/1126 [23:56<06:51,  1.28it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 53%|█████▎    | 602/1126 [24:00<15:07,  1.73s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 54%|█████▎    | 604/1126 [24:05<16:26,  1.89s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 54%|█████▎    | 605/1126 [24:09<21:37,  2.49s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 54%|█████▍    | 607/1126 [24:14<20:06,  2.32s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 54%|█████▍    | 611/1126 [24:19<11:05,  1.29s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 55%|█████▍    | 616/1126 [24:25<07:02,  1.21it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 55%|█████▍    | 619/1126 [24:31<11:21,  1.34s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 55%|█████▌    | 620/1126 [24:35<18:46,  2.23s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 55%|█████▌    | 623/1126 [24:42<16:52,  2.01s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 55%|█████▌    | 624/1126 [24:46<21:55,  2.62s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 56%|█████▌    | 625/1126 [24:50<25:19,  3.03s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 56%|█████▌    | 626/1126 [24:54<27:24,  3.29s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 56%|█████▌    | 628/1126 [24:59<23:34,  2.84s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 56%|█████▌    | 629/1126 [25:03<26:13,  3.17s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 56%|█████▋    | 635/1126 [25:11<09:02,  1.10s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 56%|█████▋    | 636/1126 [25:15<15:55,  1.95s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 57%|█████▋    | 645/1126 [25:24<05:16,  1.52it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 58%|█████▊    | 648/1126 [25:29<08:35,  1.08s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 58%|█████▊    | 651/1126 [25:34<09:15,  1.17s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 58%|█████▊    | 652/1126 [25:38<15:37,  1.98s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 58%|█████▊    | 653/1126 [25:42<20:04,  2.55s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 58%|█████▊    | 657/1126 [25:48<11:57,  1.53s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 58%|█████▊    | 658/1126 [25:52<17:55,  2.30s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 59%|█████▊    | 659/1126 [25:55<20:12,  2.60s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 59%|█████▉    | 665/1126 [26:03<07:37,  1.01it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 59%|█████▉    | 666/1126 [26:07<14:49,  1.93s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 60%|█████▉    | 670/1126 [26:12<09:10,  1.21s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 60%|█████▉    | 671/1126 [26:17<15:44,  2.08s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 60%|██████    | 680/1126 [26:30<06:33,  1.13it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 61%|██████    | 685/1126 [26:37<06:02,  1.22it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 61%|██████    | 689/1126 [26:44<08:49,  1.21s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 61%|██████▏   | 691/1126 [26:49<11:36,  1.60s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 62%|██████▏   | 694/1126 [26:54<09:37,  1.34s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 62%|██████▏   | 695/1126 [26:58<16:12,  2.26s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 62%|██████▏   | 696/1126 [27:02<20:13,  2.82s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 62%|██████▏   | 697/1126 [27:06<23:00,  3.22s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 62%|██████▏   | 698/1126 [27:10<23:38,  3.31s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 62%|██████▏   | 701/1126 [27:15<13:59,  1.98s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 62%|██████▏   | 702/1126 [27:19<18:17,  2.59s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 62%|██████▏   | 703/1126 [27:23<21:39,  3.07s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 63%|██████▎   | 704/1126 [27:27<23:29,  3.34s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 63%|██████▎   | 705/1126 [27:31<24:26,  3.48s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 63%|██████▎   | 708/1126 [27:35<13:30,  1.94s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 63%|██████▎   | 709/1126 [27:39<17:03,  2.45s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 63%|██████▎   | 710/1126 [27:43<20:51,  3.01s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 63%|██████▎   | 711/1126 [27:47<22:54,  3.31s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 63%|██████▎   | 713/1126 [27:52<18:25,  2.68s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 64%|██████▎   | 717/1126 [28:05<14:42,  2.16s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 64%|██████▍   | 721/1126 [28:10<08:23,  1.24s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 64%|██████▍   | 723/1126 [28:15<11:45,  1.75s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 64%|██████▍   | 726/1126 [28:21<10:11,  1.53s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 65%|██████▍   | 731/1126 [28:29<10:05,  1.53s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 65%|██████▌   | 732/1126 [28:33<14:35,  2.22s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 65%|██████▌   | 733/1126 [28:37<18:43,  2.86s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 65%|██████▌   | 735/1126 [28:42<16:47,  2.58s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 65%|██████▌   | 736/1126 [28:46<18:42,  2.88s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 66%|██████▌   | 739/1126 [28:51<11:35,  1.80s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 66%|██████▌   | 740/1126 [28:55<16:09,  2.51s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 66%|██████▌   | 742/1126 [28:59<14:05,  2.20s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 66%|██████▌   | 745/1126 [29:04<10:22,  1.63s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 66%|██████▋   | 746/1126 [29:08<14:58,  2.37s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 66%|██████▋   | 747/1126 [29:14<20:00,  3.17s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 67%|██████▋   | 753/1126 [29:22<08:01,  1.29s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 67%|██████▋   | 754/1126 [29:26<12:52,  2.08s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 67%|██████▋   | 755/1126 [29:30<16:20,  2.64s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 67%|██████▋   | 758/1126 [29:35<10:54,  1.78s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 67%|██████▋   | 760/1126 [29:40<11:34,  1.90s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 68%|██████▊   | 762/1126 [29:45<12:33,  2.07s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 68%|██████▊   | 764/1126 [29:51<14:02,  2.33s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 68%|██████▊   | 766/1126 [29:56<13:26,  2.24s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 68%|██████▊   | 770/1126 [30:02<08:12,  1.38s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 68%|██████▊   | 771/1126 [30:06<12:36,  2.13s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 69%|██████▊   | 772/1126 [30:10<16:38,  2.82s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 69%|██████▊   | 773/1126 [30:14<18:53,  3.21s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 69%|██████▊   | 774/1126 [30:18<20:03,  3.42s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 69%|██████▉   | 775/1126 [30:23<22:03,  3.77s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 69%|██████▉   | 776/1126 [30:28<24:46,  4.25s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 69%|██████▉   | 778/1126 [30:32<17:29,  3.02s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 69%|██████▉   | 779/1126 [30:36<18:21,  3.18s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 69%|██████▉   | 781/1126 [30:40<14:33,  2.53s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 69%|██████▉   | 782/1126 [30:44<16:29,  2.88s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 70%|██████▉   | 784/1126 [30:49<13:40,  2.40s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 70%|██████▉   | 785/1126 [30:52<16:16,  2.86s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 70%|██████▉   | 788/1126 [30:59<11:38,  2.07s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 70%|███████   | 790/1126 [31:03<11:13,  2.00s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 71%|███████   | 796/1126 [31:11<05:22,  1.02it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 71%|███████   | 798/1126 [31:15<07:49,  1.43s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 71%|███████   | 799/1126 [31:19<11:56,  2.19s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 71%|███████   | 801/1126 [31:25<12:26,  2.30s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 71%|███████   | 802/1126 [31:29<15:01,  2.78s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 71%|███████▏  | 804/1126 [31:34<13:46,  2.57s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 72%|███████▏  | 806/1126 [31:39<12:02,  2.26s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 72%|███████▏  | 809/1126 [31:45<09:37,  1.82s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 72%|███████▏  | 810/1126 [31:49<14:10,  2.69s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 73%|███████▎  | 818/1126 [32:00<04:38,  1.11it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 73%|███████▎  | 819/1126 [32:05<10:33,  2.06s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 73%|███████▎  | 824/1126 [32:14<07:02,  1.40s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 73%|███████▎  | 825/1126 [32:19<12:33,  2.50s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 73%|███████▎  | 827/1126 [32:24<11:41,  2.35s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 74%|███████▍  | 832/1126 [32:32<06:04,  1.24s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 74%|███████▍  | 833/1126 [32:37<11:58,  2.45s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 74%|███████▍  | 834/1126 [32:42<15:46,  3.24s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 74%|███████▍  | 838/1126 [32:49<08:23,  1.75s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 75%|███████▍  | 840/1126 [32:55<10:29,  2.20s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 75%|███████▍  | 842/1126 [33:02<12:41,  2.68s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 75%|███████▍  | 843/1126 [33:07<15:38,  3.32s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 75%|███████▍  | 844/1126 [33:12<17:03,  3.63s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 75%|███████▌  | 845/1126 [33:16<18:42,  3.99s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 75%|███████▌  | 847/1126 [33:21<14:25,  3.10s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 76%|███████▌  | 853/1126 [33:29<04:59,  1.10s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 76%|███████▌  | 854/1126 [33:33<09:00,  1.99s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 76%|███████▌  | 858/1126 [33:39<05:15,  1.18s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 76%|███████▋  | 859/1126 [33:43<09:05,  2.04s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 76%|███████▋  | 860/1126 [33:46<11:04,  2.50s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 76%|███████▋  | 861/1126 [33:50<12:50,  2.91s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 77%|███████▋  | 862/1126 [33:54<14:00,  3.18s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 77%|███████▋  | 867/1126 [34:01<05:18,  1.23s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 77%|███████▋  | 868/1126 [34:05<08:26,  1.96s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 77%|███████▋  | 869/1126 [34:08<10:47,  2.52s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 77%|███████▋  | 870/1126 [34:13<12:51,  3.02s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 77%|███████▋  | 871/1126 [34:17<14:21,  3.38s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 77%|███████▋  | 872/1126 [34:21<14:58,  3.54s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 78%|███████▊  | 873/1126 [34:25<15:16,  3.62s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 78%|███████▊  | 878/1126 [34:32<06:02,  1.46s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 79%|███████▊  | 884/1126 [34:39<03:08,  1.28it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 79%|███████▉  | 890/1126 [34:46<02:54,  1.35it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 79%|███████▉  | 891/1126 [34:50<07:03,  1.80s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 79%|███████▉  | 892/1126 [34:54<10:01,  2.57s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 79%|███████▉  | 895/1126 [35:00<06:52,  1.79s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 80%|███████▉  | 898/1126 [35:04<05:15,  1.38s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 80%|████████  | 903/1126 [35:11<03:24,  1.09it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 80%|████████  | 905/1126 [35:17<07:04,  1.92s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 80%|████████  | 906/1126 [35:21<09:18,  2.54s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 81%|████████  | 907/1126 [35:25<10:40,  2.93s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 81%|████████  | 908/1126 [35:29<11:56,  3.29s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 81%|████████  | 909/1126 [35:32<12:16,  3.39s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 81%|████████  | 911/1126 [35:38<11:12,  3.13s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 81%|████████  | 913/1126 [35:43<09:22,  2.64s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 81%|████████▏ | 916/1126 [35:49<06:33,  1.87s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 82%|████████▏ | 919/1126 [35:55<05:38,  1.64s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 82%|████████▏ | 924/1126 [36:01<03:29,  1.03s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 82%|████████▏ | 926/1126 [36:06<05:38,  1.69s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 82%|████████▏ | 927/1126 [36:11<08:34,  2.58s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 83%|████████▎ | 931/1126 [36:16<04:36,  1.42s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 83%|████████▎ | 932/1126 [36:20<06:48,  2.10s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 83%|████████▎ | 936/1126 [36:26<03:57,  1.25s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 83%|████████▎ | 937/1126 [36:29<06:17,  2.00s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 83%|████████▎ | 938/1126 [36:33<08:03,  2.57s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 83%|████████▎ | 939/1126 [36:37<09:25,  3.02s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 83%|████████▎ | 940/1126 [36:41<10:17,  3.32s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 84%|████████▎ | 942/1126 [36:47<08:50,  2.88s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 84%|████████▍ | 946/1126 [36:57<05:54,  1.97s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 84%|████████▍ | 947/1126 [37:03<08:50,  2.96s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 84%|████████▍ | 948/1126 [37:07<10:34,  3.56s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 84%|████████▍ | 949/1126 [37:13<11:55,  4.04s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 85%|████████▍ | 952/1126 [37:19<07:21,  2.54s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 85%|████████▍ | 953/1126 [37:24<09:17,  3.22s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 85%|████████▍ | 955/1126 [37:30<08:13,  2.88s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 85%|████████▍ | 957/1126 [37:36<07:47,  2.77s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 85%|████████▌ | 959/1126 [37:42<06:58,  2.51s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 85%|████████▌ | 960/1126 [37:45<08:06,  2.93s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 85%|████████▌ | 961/1126 [37:49<08:46,  3.19s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 86%|████████▌ | 964/1126 [37:54<05:12,  1.93s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 86%|████████▌ | 965/1126 [37:59<07:05,  2.64s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 86%|████████▌ | 969/1126 [38:05<03:58,  1.52s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 87%|████████▋ | 980/1126 [38:14<01:08,  2.13it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 87%|████████▋ | 981/1126 [38:19<03:55,  1.63s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 87%|████████▋ | 982/1126 [38:23<06:05,  2.54s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 87%|████████▋ | 984/1126 [38:28<05:39,  2.39s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 88%|████████▊ | 987/1126 [38:33<03:52,  1.67s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 88%|████████▊ | 989/1126 [38:38<04:09,  1.82s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 88%|████████▊ | 993/1126 [38:44<02:43,  1.23s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 88%|████████▊ | 994/1126 [38:48<04:30,  2.05s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 88%|████████▊ | 995/1126 [38:52<06:00,  2.75s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 88%|████████▊ | 996/1126 [38:56<06:40,  3.08s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 89%|████████▉ | 1005/1126 [39:06<01:40,  1.21it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 89%|████████▉ | 1007/1126 [39:11<02:45,  1.39s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 90%|████████▉ | 1008/1126 [39:15<04:17,  2.18s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 90%|████████▉ | 1009/1126 [39:19<05:15,  2.69s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 90%|████████▉ | 1010/1126 [39:23<06:00,  3.11s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 90%|████████▉ | 1011/1126 [39:27<06:31,  3.40s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 90%|█████████ | 1015/1126 [39:34<03:25,  1.85s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 90%|█████████ | 1016/1126 [39:38<04:30,  2.46s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 90%|█████████ | 1019/1126 [39:44<03:06,  1.75s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 91%|█████████ | 1023/1126 [39:49<01:56,  1.13s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 92%|█████████▏| 1035/1126 [40:00<00:46,  1.94it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 92%|█████████▏| 1036/1126 [40:04<02:16,  1.52s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 92%|█████████▏| 1037/1126 [40:08<03:46,  2.54s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 92%|█████████▏| 1040/1126 [40:14<02:33,  1.78s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 92%|█████████▏| 1041/1126 [40:18<03:30,  2.48s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 93%|█████████▎| 1045/1126 [40:23<01:46,  1.31s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 93%|█████████▎| 1046/1126 [40:27<02:42,  2.03s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 93%|█████████▎| 1047/1126 [40:31<03:22,  2.56s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 94%|█████████▎| 1055/1126 [40:39<00:54,  1.30it/s]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 94%|█████████▍| 1056/1126 [40:44<02:05,  1.79s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 94%|█████████▍| 1059/1126 [40:49<01:43,  1.54s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 94%|█████████▍| 1060/1126 [40:53<02:31,  2.30s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 94%|█████████▍| 1061/1126 [40:57<03:03,  2.82s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 94%|█████████▍| 1062/1126 [41:01<03:27,  3.25s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 94%|█████████▍| 1063/1126 [41:06<03:45,  3.58s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 94%|█████████▍| 1064/1126 [41:10<03:53,  3.76s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 95%|█████████▍| 1066/1126 [41:14<02:42,  2.72s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 95%|█████████▍| 1067/1126 [41:18<02:58,  3.03s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 95%|█████████▍| 1069/1126 [41:22<02:20,  2.47s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 95%|█████████▌| 1071/1126 [41:27<02:09,  2.36s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 95%|█████████▌| 1072/1126 [41:31<02:33,  2.84s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 95%|█████████▌| 1073/1126 [41:35<02:46,  3.14s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 95%|█████████▌| 1074/1126 [41:39<03:00,  3.46s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 95%|█████████▌| 1075/1126 [41:43<02:59,  3.52s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 96%|█████████▌| 1076/1126 [41:47<03:02,  3.64s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 96%|█████████▌| 1077/1126 [41:51<03:01,  3.70s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 96%|█████████▌| 1081/1126 [41:56<01:11,  1.60s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 96%|█████████▌| 1082/1126 [42:00<01:38,  2.24s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 96%|█████████▋| 1084/1126 [42:04<01:28,  2.10s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 96%|█████████▋| 1085/1126 [42:09<01:55,  2.81s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 96%|█████████▋| 1086/1126 [42:13<02:04,  3.12s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 97%|█████████▋| 1087/1126 [42:17<02:16,  3.49s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 97%|█████████▋| 1088/1126 [42:21<02:20,  3.71s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 97%|█████████▋| 1090/1126 [42:27<01:50,  3.07s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 97%|█████████▋| 1092/1126 [42:31<01:27,  2.57s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 97%|█████████▋| 1093/1126 [42:35<01:39,  3.00s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 97%|█████████▋| 1095/1126 [42:40<01:19,  2.57s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 97%|█████████▋| 1097/1126 [42:45<01:05,  2.27s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 98%|█████████▊| 1099/1126 [42:49<00:55,  2.06s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 98%|█████████▊| 1100/1126 [42:53<01:07,  2.59s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 98%|█████████▊| 1101/1126 [42:57<01:12,  2.91s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 98%|█████████▊| 1102/1126 [43:01<01:19,  3.32s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 98%|█████████▊| 1103/1126 [43:05<01:21,  3.54s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 98%|█████████▊| 1108/1126 [43:11<00:23,  1.31s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 99%|█████████▉| 1112/1126 [43:17<00:14,  1.02s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 99%|█████████▉| 1115/1126 [43:22<00:12,  1.17s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 99%|█████████▉| 1116/1126 [43:26<00:21,  2.12s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 99%|█████████▉| 1118/1126 [43:31<00:17,  2.22s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      " 99%|█████████▉| 1120/1126 [43:36<00:12,  2.10s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "100%|█████████▉| 1121/1126 [43:40<00:13,  2.63s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "100%|█████████▉| 1122/1126 [43:44<00:12,  3.02s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "100%|█████████▉| 1124/1126 [43:49<00:05,  2.67s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "100%|█████████▉| 1125/1126 [43:53<00:03,  3.17s/it]Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "100%|██████████| 1126/1126 [43:57<00:00,  2.34s/it]\n"
     ]
    }
   ],
   "source": [
    "import csv\n",
    "import random\n",
    "import pprint\n",
    "import pandas as pd\n",
    "import tqdm\n",
    "\n",
    "to_model = []\n",
    "\n",
    "with open(\"./hidden_test_only.csv\") as csv_file: #public_test_only\n",
    "    for row in csv_file.readlines():\n",
    "        source = row.strip()\n",
    "        to_model.append(source)\n",
    "\n",
    "print(len(to_model))\n",
    "\n",
    "with open(\"hypothesis_gptXL_private_subset_check_no_sample.txt\", \"w\") as out_file:\n",
    "    for sentence in tqdm.tqdm(to_model):\n",
    "        mysentence = \"<s>\" + sentence + \" ==> \"\n",
    "        \n",
    "        ress = generate_nosample(sentence)[0]\n",
    "        try:\n",
    "            result = ress.split(\"==>\")[1].replace(\"</s>\", \"\").replace(\"</s\", \"\").replace(\"<pad>\", \"\").replace(\"<pad\", \"\").strip()\n",
    "            if check_bad_generations(sentence, result):\n",
    "                result = get_paraphraser_var(sentence)\n",
    "            out_file.write(result+\"\\n\")\n",
    "        except:\n",
    "            result = get_paraphraser_var(sentence)\n",
    "            out_file.write(result+\"\\n\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### To check of two hypothessis use Sentence Transformer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: sentence-transformers in /home/user/conda/lib/python3.7/site-packages (0.4.0)\n",
      "Collecting sentence-transformers\n",
      "  Using cached sentence_transformers-1.2.0-py3-none-any.whl\n",
      "Requirement already satisfied: numpy in /home/user/conda/lib/python3.7/site-packages (from sentence-transformers) (1.18.5)\n",
      "Requirement already satisfied: tqdm in /home/user/conda/lib/python3.7/site-packages (from sentence-transformers) (4.60.0)\n",
      "Requirement already satisfied: torch>=1.6.0 in /home/user/conda/lib/python3.7/site-packages (from sentence-transformers) (1.6.0+cu101)\n",
      "Requirement already satisfied: scikit-learn in /home/user/conda/lib/python3.7/site-packages (from sentence-transformers) (0.24.1)\n",
      "Requirement already satisfied: sentencepiece in /home/user/conda/lib/python3.7/site-packages (from sentence-transformers) (0.1.95)\n",
      "Requirement already satisfied: nltk in /home/user/conda/lib/python3.7/site-packages (from sentence-transformers) (3.6.1)\n",
      "Requirement already satisfied: scipy in /home/user/conda/lib/python3.7/site-packages (from sentence-transformers) (1.4.1)\n",
      "Requirement already satisfied: transformers<5.0.0,>=3.1.0 in /home/user/conda/lib/python3.7/site-packages (from sentence-transformers) (4.6.1)\n",
      "Requirement already satisfied: torchvision in /home/user/conda/lib/python3.7/site-packages (from sentence-transformers) (0.7.0+cu101)\n",
      "Requirement already satisfied: future in /home/user/conda/lib/python3.7/site-packages (from torch>=1.6.0->sentence-transformers) (0.18.2)\n",
      "Requirement already satisfied: requests in /home/user/conda/lib/python3.7/site-packages (from transformers<5.0.0,>=3.1.0->sentence-transformers) (2.25.1)\n",
      "Requirement already satisfied: packaging in /home/user/conda/lib/python3.7/site-packages (from transformers<5.0.0,>=3.1.0->sentence-transformers) (20.9)\n",
      "Requirement already satisfied: importlib-metadata in /home/user/conda/lib/python3.7/site-packages (from transformers<5.0.0,>=3.1.0->sentence-transformers) (3.10.0)\n",
      "Requirement already satisfied: huggingface-hub==0.0.8 in /home/user/conda/lib/python3.7/site-packages (from transformers<5.0.0,>=3.1.0->sentence-transformers) (0.0.8)\n",
      "Requirement already satisfied: tokenizers<0.11,>=0.10.1 in /home/user/conda/lib/python3.7/site-packages (from transformers<5.0.0,>=3.1.0->sentence-transformers) (0.10.3)\n",
      "Requirement already satisfied: regex!=2019.12.17 in /home/user/conda/lib/python3.7/site-packages (from transformers<5.0.0,>=3.1.0->sentence-transformers) (2021.4.4)\n",
      "Requirement already satisfied: sacremoses in /home/user/conda/lib/python3.7/site-packages (from transformers<5.0.0,>=3.1.0->sentence-transformers) (0.0.45)\n",
      "Requirement already satisfied: filelock in /home/user/conda/lib/python3.7/site-packages (from transformers<5.0.0,>=3.1.0->sentence-transformers) (3.0.12)\n",
      "Requirement already satisfied: zipp>=0.5 in /home/user/conda/lib/python3.7/site-packages (from importlib-metadata->transformers<5.0.0,>=3.1.0->sentence-transformers) (3.4.1)\n",
      "Requirement already satisfied: typing-extensions>=3.6.4 in /home/user/conda/lib/python3.7/site-packages (from importlib-metadata->transformers<5.0.0,>=3.1.0->sentence-transformers) (3.7.4.3)\n",
      "Requirement already satisfied: joblib in /home/user/conda/lib/python3.7/site-packages (from nltk->sentence-transformers) (1.0.1)\n",
      "Requirement already satisfied: click in /home/user/conda/lib/python3.7/site-packages (from nltk->sentence-transformers) (7.1.2)\n",
      "Requirement already satisfied: pyparsing>=2.0.2 in /home/user/conda/lib/python3.7/site-packages (from packaging->transformers<5.0.0,>=3.1.0->sentence-transformers) (2.4.7)\n",
      "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /home/user/conda/lib/python3.7/site-packages (from requests->transformers<5.0.0,>=3.1.0->sentence-transformers) (1.26.4)\n",
      "Requirement already satisfied: chardet<5,>=3.0.2 in /home/user/conda/lib/python3.7/site-packages (from requests->transformers<5.0.0,>=3.1.0->sentence-transformers) (4.0.0)\n",
      "Requirement already satisfied: idna<3,>=2.5 in /home/user/conda/lib/python3.7/site-packages (from requests->transformers<5.0.0,>=3.1.0->sentence-transformers) (2.10)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in /home/user/conda/lib/python3.7/site-packages (from requests->transformers<5.0.0,>=3.1.0->sentence-transformers) (2020.12.5)\n",
      "Requirement already satisfied: six in /home/user/conda/lib/python3.7/site-packages (from sacremoses->transformers<5.0.0,>=3.1.0->sentence-transformers) (1.15.0)\n",
      "Requirement already satisfied: threadpoolctl>=2.0.0 in /home/user/conda/lib/python3.7/site-packages (from scikit-learn->sentence-transformers) (2.1.0)\n",
      "Requirement already satisfied: pillow>=4.1.1 in /home/user/conda/lib/python3.7/site-packages (from torchvision->sentence-transformers) (8.2.0)\n",
      "Installing collected packages: sentence-transformers\n",
      "  Attempting uninstall: sentence-transformers\n",
      "    Found existing installation: sentence-transformers 0.4.0\n",
      "    Uninstalling sentence-transformers-0.4.0:\n",
      "      Successfully uninstalled sentence-transformers-0.4.0\n",
      "\u001b[31mERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.\n",
      "russian-paraphrasers 0.0.3 requires sentence-transformers==0.4.0, but you have sentence-transformers 1.2.0 which is incompatible.\u001b[0m\n",
      "Successfully installed sentence-transformers-1.2.0\n"
     ]
    }
   ],
   "source": [
    "!pip install -U sentence-transformers\n",
    "\n",
    "\n",
    "from sentence_transformers import SentenceTransformer, util\n",
    "model = SentenceTransformer('paraphrase-xlm-r-multilingual-v1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[0.6543]]) tensor([[0.8849]])\n",
      "['Роскомнадзор неоднократно привлекал Facebook к административной ответственности по постановлению мирового судьи.']\n"
     ]
    }
   ],
   "source": [
    "public_test = ['Facebook по инициативе Роскомнадзора неоднократно привлекалась к административной ответственности постановлениями мировых судей судебных участков района \"Таганский\" города Москвы.']\n",
    "data_preds_1 = [\"Роскомнадзор привлекался к ответственности за нарушение Facebook авторских прав.\"]\n",
    "data_preds_2 = [\"Роскомнадзор неоднократно привлекал Facebook к административной ответственности по постановлению мирового судьи.\"]\n",
    "\n",
    "\n",
    "best_pred = []\n",
    "first = 0\n",
    "for origin, pr1, pr2 in zip(public_test, data_preds_1, data_preds_2):\n",
    "\n",
    "    or_embedding = model.encode(origin)\n",
    "    pr1_embedding = model.encode(pr1)\n",
    "    pr2_embedding = model.encode(pr2)\n",
    "\n",
    "    op1 = util.pytorch_cos_sim(or_embedding, pr1_embedding)\n",
    "    op2 = util.pytorch_cos_sim(or_embedding, pr2_embedding)\n",
    "    print(op1, op2)\n",
    "    \n",
    "    if op1 > op2:\n",
    "        best_pred.append(pr1)\n",
    "        first += 1\n",
    "    else:\n",
    "        best_pred.append(pr2)\n",
    "\n",
    "print(best_pred)"
   ]
  },
  {
   "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": "3.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}