简介
量化涉及将模型的权重和激活值从float转换为int,这可以使模型大小跟小和推理的速度更快,而只损失很小的精度。
使用PyTorch示例中的单词语言模型(word language model),将最简单的量化形式,动态量化(dynamic quantization)应用于基于LSTM的下一个单词预测模型。
# imports
import os
from io import open
import time
import torch
import torch.nn as nn
import torch.nn.functional as F
定义模型
按照单词语言模型示例中的模型(model)定义LSTM模型架构。
class LSTMModel(nn.Module):
"""Container module with an encoder, a recurrent module, and a decoder."""
def __init__(self, ntoken, ninp, nhid, nlayers, dropout=0.5):
super(LSTMModel, self).__init__()
self.drop = nn.Dropout(dropout)
self.encoder = nn.Embedding(ntoken, ninp)
self.rnn = nn.LSTM(ninp, nhid, nlayers, dropout=dropout)
self.decoder = nn.Linear(nhid, ntoken)
self.init_weights()
self.nhid = nhid
self.nlayers = nlayers
def init_weights(self):
initrange = 0.1
self.encoder.weight.data.uniform_(-initrange, initrange)
self.decoder.bias.data.zero_()
self.decoder.weight.data.uniform_(-initrange, initrange)
def forward(self, input, hidden):
emb = self.drop(self.encoder(input))
output, hidden = self.rnn(emb, hidden)
output = self.drop(output)
decoded = self.decoder(output)
return decoded, hidden
def init_hidden(self, bsz):
weight = next(self.parameters())
return (weight.new_zeros(self.nlayers, bsz, self.nhid),
weight.new_zeros(self.nlayers, bsz, self.nhid))
加载文本数据
接下来,再次按照单词语言模型示例的预处理(preprocessing)过程,将Wikitext-2数据集(Wikitext-2 dataset)加载到语料库中。
class Dictionary(object):
def __init__(self):
self.word2idx = {}
self.idx2word = []
def add_word(self, word):
if word not in self.word2idx:
self.idx2word.append(word)
self.word2idx[word] = len(self.idx2word) - 1
return self.word2idx[word]
def __len__(self):
return len(self.idx2word)
class Corpus(object):
def __init__(self, path):
self.dictionary = Dictionary()
self.train = self.tokenize(os.path.join(path, 'train.txt'))
self.valid = self.tokenize(os.path.join(path, 'valid.txt'))
self.test = self.tokenize(os.path.join(path, 'test.txt'))
def tokenize(self, path):
"""Tokenizes a text file."""
assert os.path.exists(path)
# Add words to the dictionary
with open(path, 'r', encoding="utf8") as f:
for line in f:
words = line.split() + ['<eos>']
for word in words:
self.dictionary.add_word(word)
# Tokenize file content
with open(path, 'r', encoding="utf8") as f:
idss = []
for line in f:
words = line.split() + ['<eos>']
ids = []
for word in words:
ids.append(self.dictionary.word2idx[word])
idss.append(torch.tensor(ids).type(torch.int64))
ids = torch.cat(idss)
return ids
model_data_filepath = 'data/'
corpus = Corpus(model_data_filepath + 'wikitext-2')
加载预训练模型
动态量化是一种在模型训练后应用的量化技术。因此,将一些预训练权重加载到该模型架构中;这些权重是在单词语言模型示例中使用默认设置进行5个epoch的训练得到的。
ntokens = len(corpus.dictionary)
model = LSTMModel(
ntoken = ntokens,
ninp = 512,
nhid = 256,
nlayers = 5,
)
model.load_state_dict(
torch.load(
model_data_filepath + 'word_language_model_quantize.pth',
map_location=torch.device('cpu')
)
)
model.eval()
print(model)
生成一些文本以确保预训练模型正常工作(here)
input_ = torch.randint(ntokens, (1, 1), dtype=torch.long)
hidden = model.init_hidden(1)
temperature = 1.0
num_words = 1000
with open(model_data_filepath + 'out.txt', 'w') as outf:
with torch.no_grad(): # no tracking history
for i in range(num_words):
output, hidden = model(input_, hidden)
word_weights = output.squeeze().div(temperature).exp().cpu()
word_idx = torch.multinomial(word_weights, 1)[0]
input_.fill_(word_idx)
word = corpus.dictionary.idx2word[word_idx]
outf.write(str(word.encode('utf-8')) + ('\n' if i % 20 == 19 else ' '))
if i % 100 == 0:
print('| Generated {}/{} words'.format(i, 1000))
with open(model_data_filepath + 'out.txt', 'r') as outf:
all_output = outf.read()
print(all_output)
只需要再定义几个辅助函数:
bptt = 25
criterion = nn.CrossEntropyLoss()
eval_batch_size = 1
# create test data set
def batchify(data, bsz):
# Work out how cleanly we can divide the dataset into bsz parts.
nbatch = data.size(0) // bsz
# Trim off any extra elements that wouldn't cleanly fit (remainders).
data = data.narrow(0, 0, nbatch * bsz)
# Evenly divide the data across the bsz batches.
return data.view(bsz, -1).t().contiguous()
test_data = batchify(corpus.test, eval_batch_size)
# Evaluation functions
def get_batch(source, i):
seq_len = min(bptt, len(source) - 1 - i)
data = source[i:i+seq_len]
target = source[i+1:i+1+seq_len].reshape(-1)
return data, target
def repackage_hidden(h):
"""Wraps hidden states in new Tensors, to detach them from their history."""
if isinstance(h, torch.Tensor):
return h.detach()
else:
return tuple(repackage_hidden(v) for v in h)
def evaluate(model_, data_source):
# Turn on evaluation mode which disables dropout.
model_.eval()
total_loss = 0.
hidden = model_.init_hidden(eval_batch_size)
with torch.no_grad():
for i in range(0, data_source.size(0) - 1, bptt):
data, targets = get_batch(data_source, i)
output, hidden = model_(data, hidden)
hidden = repackage_hidden(hidden)
output_flat = output.view(-1, ntokens)
total_loss += len(data) * criterion(output_flat, targets).item()
return total_loss / (len(data_source) - 1)
测试动态量化
最后,可以在模型上调用torch.quantization.quantize_dynamic,具体地:
- 在模型中指定nn.LSTM和nn.Linear模块被量化
- 指定将权重转换为int8值
import torch.quantization
quantized_model = torch.quantization.quantize_dynamic(
model, {nn.LSTM, nn.Linear}, dtype=torch.qint8
)
print(quantized_model)
量化前后模型输出的区别
量化前:
LSTMModel(
(drop): Dropout(p=0.5, inplace=False)
(encoder): Embedding(33278, 512)
(rnn): LSTM(512, 256, num_layers=5, dropout=0.5)
(decoder): Linear(in_features=256, out_features=33278, bias=True)
)
量化后:
LSTMModel(
(drop): Dropout(p=0.5, inplace=False)
(encoder): Embedding(33278, 512)
(rnn): DynamicQuantizedLSTM(512, 256, num_layers=5, dropout=0.5)
(decoder): DynamicQuantizedLinear(in_features=256, out_features=33278, dtype=torch.qint8, qscheme=torch.per_tensor_affine)
)
首先,看到模型大小显著减少:
def print_size_of_model(model):
torch.save(model.state_dict(), "temp.p")
print('Size (MB):', os.path.getsize("temp.p")/1e6)
os.remove('temp.p')
print_size_of_model(model)
print_size_of_model(quantized_model)
量化前后模型大小,量化前Size (MB): 113.943637,量化后Size (MB): 79.738057
其次,看到推理时间更快,评估损失没有差异。对于单线程比较,将线程数设置为1,因为量化模型单线程运行。
torch.set_num_threads(1)
def time_model_evaluation(model, test_data):
s = time.time()
loss = evaluate(model, test_data)
elapsed = time.time() - s
print('''loss: {0:.3f}\nelapsed time (seconds): {1:.1f}'''.format(loss, elapsed))
time_model_evaluation(model, test_data)
time_model_evaluation(quantized_model, test_data)
输出结果
loss: 5.167
elapsed time (seconds): 211.1
loss: 5.168
elapsed time (seconds): 111.0
原文链接
https://pytorch.org/tutorials/advanced/dynamic_quantization_tutorial.html
https://github.com/pytorch/examples/tree/main/word_language_model