{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# GPU Environment"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-info\">\n",
    "\n",
    "This tutorial is available as an IPython notebook at [Malaya/example/gpu-environment](https://github.com/huseinzol05/Malaya/tree/master/example/gpu-environment).\n",
    "    \n",
    "</div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 2.9 s, sys: 3.7 s, total: 6.6 s\n",
      "Wall time: 2.09 s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/husein/dev/malaya/malaya/tokenizer.py:214: FutureWarning: Possible nested set at position 3397\n",
      "  self.tok = re.compile(r'({})'.format('|'.join(pipeline)))\n",
      "/home/husein/dev/malaya/malaya/tokenizer.py:214: FutureWarning: Possible nested set at position 3927\n",
      "  self.tok = re.compile(r'({})'.format('|'.join(pipeline)))\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "import malaya\n",
    "import logging\n",
    "logging.basicConfig(level = logging.INFO)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### List available GPU\n",
    "\n",
    "**You must install Pytorch GPU version first to enable GPU hardware acceleration**."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import torch\n",
    "\n",
    "torch.cuda.device_count()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Run model inside GPU\n",
    "\n",
    "Once you initiate `cuda` method from pytorch object, all inputs will auto cast to `cuda`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'mesolitica/translation-t5-tiny-standard-bahasa-cased': {'Size (MB)': 139,\n",
       "  'Suggested length': 1536,\n",
       "  'en-ms chrF2++': 65.91,\n",
       "  'ms-en chrF2++': 61.3,\n",
       "  'ind-ms chrF2++': 58.15,\n",
       "  'jav-ms chrF2++': 49.33,\n",
       "  'pasar ms-ms chrF2++': 58.46,\n",
       "  'pasar ms-en chrF2++': 55.76,\n",
       "  'manglish-ms chrF2++': 51.04,\n",
       "  'manglish-en chrF2++': 52.2,\n",
       "  'from lang': ['en', 'ms', 'ind', 'jav', 'bjn', 'manglish', 'pasar ms'],\n",
       "  'to lang': ['en', 'ms']},\n",
       " 'mesolitica/translation-t5-small-standard-bahasa-cased': {'Size (MB)': 242,\n",
       "  'Suggested length': 1536,\n",
       "  'en-ms chrF2++': 67.37,\n",
       "  'ms-en chrF2++': 63.79,\n",
       "  'ind-ms chrF2++': 58.09,\n",
       "  'jav-ms chrF2++': 52.11,\n",
       "  'pasar ms-ms chrF2++': 62.49,\n",
       "  'pasar ms-en chrF2++': 60.77,\n",
       "  'manglish-ms chrF2++': 52.84,\n",
       "  'manglish-en chrF2++': 53.65,\n",
       "  'from lang': ['en', 'ms', 'ind', 'jav', 'bjn', 'manglish', 'pasar ms'],\n",
       "  'to lang': ['en', 'ms']},\n",
       " 'mesolitica/translation-t5-base-standard-bahasa-cased': {'Size (MB)': 892,\n",
       "  'Suggested length': 1536,\n",
       "  'en-ms chrF2++': 67.62,\n",
       "  'ms-en chrF2++': 64.41,\n",
       "  'ind-ms chrF2++': 59.25,\n",
       "  'jav-ms chrF2++': 52.86,\n",
       "  'pasar ms-ms chrF2++': 62.99,\n",
       "  'pasar ms-en chrF2++': 62.06,\n",
       "  'manglish-ms chrF2++': 54.4,\n",
       "  'manglish-en chrF2++': 54.14,\n",
       "  'from lang': ['en', 'ms', 'ind', 'jav', 'bjn', 'manglish', 'pasar ms'],\n",
       "  'to lang': ['en', 'ms']},\n",
       " 'mesolitica/translation-nanot5-tiny-malaysian-cased': {'Size (MB)': 205,\n",
       "  'Suggested length': 2048,\n",
       "  'en-ms chrF2++': 63.61,\n",
       "  'ms-en chrF2++': 59.55,\n",
       "  'ind-ms chrF2++': 56.38,\n",
       "  'jav-ms chrF2++': 47.68,\n",
       "  'mandarin-ms chrF2++': 36.61,\n",
       "  'mandarin-en chrF2++': 39.78,\n",
       "  'pasar ms-ms chrF2++': 58.74,\n",
       "  'pasar ms-en chrF2++': 54.87,\n",
       "  'manglish-ms chrF2++': 50.76,\n",
       "  'manglish-en chrF2++': 53.16,\n",
       "  'from lang': ['en',\n",
       "   'ms',\n",
       "   'ind',\n",
       "   'jav',\n",
       "   'bjn',\n",
       "   'manglish',\n",
       "   'pasar ms',\n",
       "   'mandarin',\n",
       "   'pasar mandarin'],\n",
       "  'to lang': ['en', 'ms']},\n",
       " 'mesolitica/translation-nanot5-small-malaysian-cased': {'Size (MB)': 358,\n",
       "  'Suggested length': 2048,\n",
       "  'en-ms chrF2++': 66.98,\n",
       "  'ms-en chrF2++': 63.52,\n",
       "  'ind-ms chrF2++': 58.1,\n",
       "  'jav-ms chrF2++': 51.55,\n",
       "  'mandarin-ms chrF2++': 46.09,\n",
       "  'mandarin-en chrF2++': 44.13,\n",
       "  'pasar ms-ms chrF2++': 63.2,\n",
       "  'pasar ms-en chrF2++': 59.78,\n",
       "  'manglish-ms chrF2++': 54.09,\n",
       "  'manglish-en chrF2++': 55.27,\n",
       "  'from lang': ['en',\n",
       "   'ms',\n",
       "   'ind',\n",
       "   'jav',\n",
       "   'bjn',\n",
       "   'manglish',\n",
       "   'pasar ms',\n",
       "   'mandarin',\n",
       "   'pasar mandarin'],\n",
       "  'to lang': ['en', 'ms']},\n",
       " 'mesolitica/translation-nanot5-base-malaysian-cased': {'Size (MB)': 990,\n",
       "  'Suggested length': 2048,\n",
       "  'en-ms chrF2++': 67.87,\n",
       "  'ms-en chrF2++': 64.79,\n",
       "  'ind-ms chrF2++': 56.98,\n",
       "  'jav-ms chrF2++': 51.21,\n",
       "  'mandarin-ms chrF2++': 47.39,\n",
       "  'mandarin-en chrF2++': 48.78,\n",
       "  'pasar ms-ms chrF2++': 65.06,\n",
       "  'pasar ms-en chrF2++': 64.03,\n",
       "  'manglish-ms chrF2++': 57.91,\n",
       "  'manglish-en chrF2++': 55.66,\n",
       "  'from lang': ['en',\n",
       "   'ms',\n",
       "   'ind',\n",
       "   'jav',\n",
       "   'bjn',\n",
       "   'manglish',\n",
       "   'pasar ms',\n",
       "   'mandarin',\n",
       "   'pasar mandarin'],\n",
       "  'to lang': ['en', 'ms']}}"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "malaya.translation.available_huggingface"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Loading the tokenizer from the `special_tokens_map.json` and the `added_tokens.json` will be removed in `transformers 5`,  it is kept for forward compatibility, but it is recommended to update your `tokenizer_config.json` by uploading it again. You will see the new `added_tokens_decoder` attribute that will store the relevant information.\n",
      "You are using the default legacy behaviour of the <class 'transformers.models.t5.tokenization_t5.T5Tokenizer'>. If you see this, DO NOT PANIC! This is expected, and simply means that the `legacy` (previous) behavior will be used so nothing changes for you. If you want to use the new behaviour, set `legacy=False`. This should only be set if you understand what it means, and thouroughly read the reason why this was added as explained in https://github.com/huggingface/transformers/pull/24565\n",
      "Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f6a79ba00cb241aca683f668f1f27497",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Downloading (…)neration_config.json:   0%|          | 0.00/164 [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = malaya.translation.huggingface(model = 'mesolitica/translation-t5-tiny-standard-bahasa-cased')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "T5ForConditionalGeneration(\n",
       "  (shared): Embedding(32103, 384)\n",
       "  (encoder): T5Stack(\n",
       "    (embed_tokens): Embedding(32103, 384)\n",
       "    (block): ModuleList(\n",
       "      (0): T5Block(\n",
       "        (layer): ModuleList(\n",
       "          (0): T5LayerSelfAttention(\n",
       "            (SelfAttention): T5Attention(\n",
       "              (q): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (k): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (v): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (o): Linear(in_features=768, out_features=384, bias=False)\n",
       "              (relative_attention_bias): Embedding(32, 12)\n",
       "            )\n",
       "            (layer_norm): T5LayerNorm()\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "          (1): T5LayerFF(\n",
       "            (DenseReluDense): T5DenseActDense(\n",
       "              (wi): Linear(in_features=384, out_features=1344, bias=False)\n",
       "              (wo): Linear(in_features=1344, out_features=384, bias=False)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "              (act): ReLU()\n",
       "            )\n",
       "            (layer_norm): T5LayerNorm()\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "      )\n",
       "      (1-3): 3 x T5Block(\n",
       "        (layer): ModuleList(\n",
       "          (0): T5LayerSelfAttention(\n",
       "            (SelfAttention): T5Attention(\n",
       "              (q): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (k): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (v): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (o): Linear(in_features=768, out_features=384, bias=False)\n",
       "            )\n",
       "            (layer_norm): T5LayerNorm()\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "          (1): T5LayerFF(\n",
       "            (DenseReluDense): T5DenseActDense(\n",
       "              (wi): Linear(in_features=384, out_features=1344, bias=False)\n",
       "              (wo): Linear(in_features=1344, out_features=384, bias=False)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "              (act): ReLU()\n",
       "            )\n",
       "            (layer_norm): T5LayerNorm()\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "      )\n",
       "    )\n",
       "    (final_layer_norm): T5LayerNorm()\n",
       "    (dropout): Dropout(p=0.1, inplace=False)\n",
       "  )\n",
       "  (decoder): T5Stack(\n",
       "    (embed_tokens): Embedding(32103, 384)\n",
       "    (block): ModuleList(\n",
       "      (0): T5Block(\n",
       "        (layer): ModuleList(\n",
       "          (0): T5LayerSelfAttention(\n",
       "            (SelfAttention): T5Attention(\n",
       "              (q): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (k): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (v): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (o): Linear(in_features=768, out_features=384, bias=False)\n",
       "              (relative_attention_bias): Embedding(32, 12)\n",
       "            )\n",
       "            (layer_norm): T5LayerNorm()\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "          (1): T5LayerCrossAttention(\n",
       "            (EncDecAttention): T5Attention(\n",
       "              (q): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (k): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (v): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (o): Linear(in_features=768, out_features=384, bias=False)\n",
       "            )\n",
       "            (layer_norm): T5LayerNorm()\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "          (2): T5LayerFF(\n",
       "            (DenseReluDense): T5DenseActDense(\n",
       "              (wi): Linear(in_features=384, out_features=1344, bias=False)\n",
       "              (wo): Linear(in_features=1344, out_features=384, bias=False)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "              (act): ReLU()\n",
       "            )\n",
       "            (layer_norm): T5LayerNorm()\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "      )\n",
       "      (1-3): 3 x T5Block(\n",
       "        (layer): ModuleList(\n",
       "          (0): T5LayerSelfAttention(\n",
       "            (SelfAttention): T5Attention(\n",
       "              (q): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (k): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (v): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (o): Linear(in_features=768, out_features=384, bias=False)\n",
       "            )\n",
       "            (layer_norm): T5LayerNorm()\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "          (1): T5LayerCrossAttention(\n",
       "            (EncDecAttention): T5Attention(\n",
       "              (q): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (k): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (v): Linear(in_features=384, out_features=768, bias=False)\n",
       "              (o): Linear(in_features=768, out_features=384, bias=False)\n",
       "            )\n",
       "            (layer_norm): T5LayerNorm()\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "          (2): T5LayerFF(\n",
       "            (DenseReluDense): T5DenseActDense(\n",
       "              (wi): Linear(in_features=384, out_features=1344, bias=False)\n",
       "              (wo): Linear(in_features=1344, out_features=384, bias=False)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "              (act): ReLU()\n",
       "            )\n",
       "            (layer_norm): T5LayerNorm()\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "      )\n",
       "    )\n",
       "    (final_layer_norm): T5LayerNorm()\n",
       "    (dropout): Dropout(p=0.1, inplace=False)\n",
       "  )\n",
       "  (lm_head): Linear(in_features=384, out_features=32103, bias=False)\n",
       ")"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.cuda()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.generate(['i like chicken'])"
   ]
  }
 ],
 "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"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}