{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Different Precision"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"\n",
"\n",
"This tutorial is available as an IPython notebook at [Malaya/example/different-precision](https://github.com/huseinzol05/Malaya/tree/master/example/different-precision).\n",
" \n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Read more at https://huggingface.co/docs/diffusers/optimization/fp16#half-precision-weights"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"CPU times: user 2.88 s, sys: 3.46 s, total: 6.34 s\n",
"Wall time: 2.21 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": "code",
"execution_count": 2,
"metadata": {
"scrolled": true
},
"outputs": [],
"source": [
"import torch"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [],
"source": [
"# https://discuss.pytorch.org/t/finding-model-size/130275\n",
"\n",
"def get_model_size_mb(model):\n",
" param_size = 0\n",
" for param in model.model.parameters():\n",
" param_size += param.nelement() * param.element_size()\n",
" buffer_size = 0\n",
" for buffer in model.model.buffers():\n",
" buffer_size += buffer.nelement() * buffer.element_size()\n",
" return (param_size + buffer_size) / 1024**2"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Load default precision, FP32"
]
},
{
"cell_type": "code",
"execution_count": 5,
"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 . 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"
]
}
],
"source": [
"model = malaya.translation.huggingface(model = 'mesolitica/translation-t5-small-standard-bahasa-cased')"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"230.765625"
]
},
"execution_count": 6,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"get_model_size_mb(model)"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/home/husein/.local/lib/python3.8/site-packages/transformers/generation/utils.py:1260: UserWarning: Using the model-agnostic default `max_length` (=20) to control the generation length. We recommend setting `max_new_tokens` to control the maximum length of the generation.\n",
" warnings.warn(\n"
]
},
{
"data": {
"text/plain": [
"['Saya suka ayam']"
]
},
"execution_count": 7,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"model.generate(['i like chicken'])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Load FP16\n",
"\n",
"**Only worked on GPU**."
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.\n"
]
}
],
"source": [
"model = malaya.translation.huggingface(model = 'mesolitica/translation-t5-small-standard-bahasa-cased',\n",
" torch_dtype=torch.float16)"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"139.3828125"
]
},
"execution_count": 10,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"get_model_size_mb(model)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Load INT8\n",
"\n",
"Required latest version `accelerate` and `bitsandbytes`,\n",
"\n",
"```bash\n",
"pip3 install accelerate bitsandbytes\n",
"```\n",
"\n",
"**Only worked on GPU**."
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.\n"
]
}
],
"source": [
"model = malaya.translation.huggingface(model = 'mesolitica/translation-t5-small-standard-bahasa-cased',\n",
" load_in_8bit=True, device_map='auto')"
]
},
{
"cell_type": "code",
"execution_count": 13,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"109.3828125"
]
},
"execution_count": 13,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"get_model_size_mb(model)"
]
},
{
"cell_type": "code",
"execution_count": 14,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"['Saya suka ayam']"
]
},
"execution_count": 14,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"model.generate(['i like chicken'])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Load INT4\n",
"\n",
"Required latest version `accelerate` and `bitsandbytes`,\n",
"\n",
"```bash\n",
"pip3 install accelerate bitsandbytes\n",
"```\n",
"\n",
"**Only worked on GPU**."
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.\n"
]
}
],
"source": [
"model = malaya.translation.huggingface(model = 'mesolitica/translation-t5-small-standard-bahasa-cased',\n",
" load_in_4bit=True, device_map='auto')"
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"94.3828125"
]
},
"execution_count": 16,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"get_model_size_mb(model)"
]
},
{
"cell_type": "code",
"execution_count": 17,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"['Saya suka ayam']"
]
},
"execution_count": 17,
"metadata": {},
"output_type": "execute_result"
}
],
"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
}