{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# MS to JAV HuggingFace"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-info\">\n",
    "\n",
    "This tutorial is available as an IPython notebook at [Malaya/example/ms-jav-translation-huggingface](https://github.com/huseinzol05/Malaya/tree/master/example/ms-jav-translation-huggingface).\n",
    "    \n",
    "</div>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-warning\">\n",
    "\n",
    "This module trained on standard language and augmented local language structures, proceed with caution.\n",
    "    \n",
    "</div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "\n",
    "os.environ['CUDA_VISIBLE_DEVICES'] = ''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 3.94 s, sys: 3.32 s, total: 7.26 s\n",
      "Wall time: 3.1 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "import malaya\n",
    "import logging\n",
    "\n",
    "logging.basicConfig(level=logging.INFO)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### List available HuggingFace models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:malaya.translation.ms_jav:tested on FLORES200 MS-JAV (zsm_Latn-jav_Latn) pair `dev` set, https://github.com/facebookresearch/flores/tree/main/flores200\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Size (MB)</th>\n",
       "      <th>BLEU</th>\n",
       "      <th>SacreBLEU Verbose</th>\n",
       "      <th>SacreBLEU-chrF++-FLORES200</th>\n",
       "      <th>Suggested length</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>mesolitica/finetune-translation-austronesian-t5-tiny-standard-bahasa-cased</th>\n",
       "      <td>139</td>\n",
       "      <td>23.796499</td>\n",
       "      <td>59.2/31.6/18.2/10.8 (BP = 0.966 ratio = 0.967 ...</td>\n",
       "      <td>51.21</td>\n",
       "      <td>512</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mesolitica/finetune-translation-austronesian-t5-small-standard-bahasa-cased</th>\n",
       "      <td>242</td>\n",
       "      <td>24.599989</td>\n",
       "      <td>58.3/31.6/18.5/11.2 (BP = 0.990 ratio = 0.990 ...</td>\n",
       "      <td>51.65</td>\n",
       "      <td>512</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mesolitica/finetune-translation-austronesian-t5-base-standard-bahasa-cased</th>\n",
       "      <td>892</td>\n",
       "      <td>24.642363</td>\n",
       "      <td>60.1/32.7/19.1/11.5 (BP = 0.961 ratio = 0.961 ...</td>\n",
       "      <td>51.91</td>\n",
       "      <td>512</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                   Size (MB)       BLEU  \\\n",
       "mesolitica/finetune-translation-austronesian-t5...       139  23.796499   \n",
       "mesolitica/finetune-translation-austronesian-t5...       242  24.599989   \n",
       "mesolitica/finetune-translation-austronesian-t5...       892  24.642363   \n",
       "\n",
       "                                                                                    SacreBLEU Verbose  \\\n",
       "mesolitica/finetune-translation-austronesian-t5...  59.2/31.6/18.2/10.8 (BP = 0.966 ratio = 0.967 ...   \n",
       "mesolitica/finetune-translation-austronesian-t5...  58.3/31.6/18.5/11.2 (BP = 0.990 ratio = 0.990 ...   \n",
       "mesolitica/finetune-translation-austronesian-t5...  60.1/32.7/19.1/11.5 (BP = 0.961 ratio = 0.961 ...   \n",
       "\n",
       "                                                   SacreBLEU-chrF++-FLORES200  \\\n",
       "mesolitica/finetune-translation-austronesian-t5...                      51.21   \n",
       "mesolitica/finetune-translation-austronesian-t5...                      51.65   \n",
       "mesolitica/finetune-translation-austronesian-t5...                      51.91   \n",
       "\n",
       "                                                   Suggested length  \n",
       "mesolitica/finetune-translation-austronesian-t5...              512  \n",
       "mesolitica/finetune-translation-austronesian-t5...              512  \n",
       "mesolitica/finetune-translation-austronesian-t5...              512  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "malaya.translation.ms_jav.available_huggingface()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load Transformer models\n",
    "\n",
    "```python\n",
    "def huggingface(\n",
    "    model: str = 'mesolitica/finetune-translation-austronesian-t5-small-standard-bahasa-cased',\n",
    "    force_check: bool = True,\n",
    "    **kwargs,\n",
    "):\n",
    "    \"\"\"\n",
    "    Load HuggingFace model to translate MS-to-JAV.\n",
    "\n",
    "    Parameters\n",
    "    ----------\n",
    "    model: str, optional (default='mesolitica/finetune-translation-t5-small-standard-bahasa-cased')\n",
    "        Check available models at `malaya.translation.ms_jav.available_huggingface()`.\n",
    "\n",
    "    Returns\n",
    "    -------\n",
    "    result: malaya.torch_model.huggingface.Generator\n",
    "    \"\"\"\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "transformer_huggingface = malaya.translation.ms_jav.huggingface()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Translate\n",
    "\n",
    "```python\n",
    "def generate(self, strings: List[str], **kwargs):\n",
    "    \"\"\"\n",
    "    Generate texts from the input.\n",
    "\n",
    "    Parameters\n",
    "    ----------\n",
    "    strings : List[str]\n",
    "    **kwargs: vector arguments pass to huggingface `generate` method.\n",
    "        Read more at https://huggingface.co/docs/transformers/main_classes/text_generation\n",
    "\n",
    "    Returns\n",
    "    -------\n",
    "    result: List[str]\n",
    "    \"\"\"\n",
    "```\n",
    "\n",
    "**For better results, always split by end of sentences**."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pprint import pprint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "text = 'saya tak suka ayam goreng dan itik'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "You're using a T5TokenizerFast tokenizer. Please note that with a fast tokenizer, using the `__call__` method is faster than using a method to encode the text followed by a call to the `pad` method to get a padded encoding.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Aku ora seneng karo pitik goreng lan bebek']\n",
      "CPU times: user 2.29 s, sys: 0 ns, total: 2.29 s\n",
      "Wall time: 198 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "pprint(transformer_huggingface.generate([text],\n",
    "                                       max_length = 1000))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.10"
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}