上一篇讲的是怎么用Trainer API来进行训练的，这一篇讲怎么用PyTorch写训练：
!pip install datasets transformers[sentencepiece]
from datasets import load_dataset
from transformers import AutoTokenizer, DataCollatorWithPadding
raw_datasets = load_dataset("glue", "mrpc")
checkpoint = "bert-base-uncased"
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
def tokenize_function(example):
return tokenizer(example["sentence1"], example["sentence2"], truncation=True)
tokenized_datasets = raw_datasets.map(tokenize_function, batched=True)
data_collator = DataCollatorWithPadding(tokenizer=tokenizer)
Downloading: 0%| | 0.00/7.78k [00:00<?, ?B/s]
Downloading: 0%| | 0.00/4.47k [00:00<?, ?B/s]
Downloading and preparing dataset glue/mrpc (download: 1.43 MiB, generated: 1.43 MiB, post-processed: Unknown size, total: 2.85 MiB) to /root/.cache/huggingface/datasets/glue/mrpc/1.0.0/dacbe3125aa31d7f70367a07a8a9e72a5a0bfeb5fc42e75c9db75b96da6053ad...
Downloading: 0.00B [00:00, ?B/s]
Downloading: 0.00B [00:00, ?B/s]
Downloading: 0.00B [00:00, ?B/s]
0 examples [00:00, ? examples/s]
0 examples [00:00, ? examples/s]
0 examples [00:00, ? examples/s]
Dataset glue downloaded and prepared to /root/.cache/huggingface/datasets/glue/mrpc/1.0.0/dacbe3125aa31d7f70367a07a8a9e72a5a0bfeb5fc42e75c9db75b96da6053ad. Subsequent calls will reuse this data.
Downloading: 0%| | 0.00/28.0 [00:00<?, ?B/s]
Downloading: 0%| | 0.00/570 [00:00<?, ?B/s]
Downloading: 0%| | 0.00/232k [00:00<?, ?B/s]
Downloading: 0%| | 0.00/466k [00:00<?, ?B/s]
0%| | 0/4 [00:00<?, ?ba/s]
0%| | 0/1 [00:00<?, ?ba/s]
0%| | 0/2 [00:00<?, ?ba/s]
Prepare for training
Postprocessing:
- Remove the columns corresponding to values the model does not expect (like the sentence1 and sentence2 columns).
- Rename the column label to labels (because the model expects the argument to be named labels).
- Set the format of the datasets so they return PyTorch tensors instead of lists.
# remove columns
tokenized_datasets = tokenized_datasets.remove_columns(
["sentence1", "sentence2", "idx"]
)
# rename label
tokenized_datasets = tokenized_datasets.rename_column("label", "labels")
# turn into tensors
tokenized_datasets.set_format("torch")
tokenized_datasets["train"].column_names
['attention_mask', 'input_ids', 'labels', 'token_type_ids']
接下来就可以定义DataLoader类了：
from torch.utils.data import DataLoader
train_dataloader = DataLoader(
tokenized_datasets["train"], shuffle=True, batch_size=8, collate_fn=data_collator
)
eval_dataloader = DataLoader(
tokenized_datasets["validation"], batch_size=8, collate_fn=data_collator
)
看一下一个batch：
for batch in train_dataloader:
break
{k: v.shape for k, v in batch.items()}
{'attention_mask': torch.Size([8, 65]),
'input_ids': torch.Size([8, 65]),
'labels': torch.Size([8]),
'token_type_ids': torch.Size([8, 65])}
数据处理彻底结束了，下面定义模型：
from transformers import AutoModelForSequenceClassification
model = AutoModelForSequenceClassification.from_pretrained(checkpoint, num_labels=2)
Downloading: 0%| | 0.00/440M [00:00<?, ?B/s]
Some weights of the model checkpoint at bert-base-uncased were not used when initializing BertForSequenceClassification: ['cls.predictions.decoder.weight', 'cls.predictions.transform.dense.weight', 'cls.predictions.bias', 'cls.predictions.transform.LayerNorm.bias', 'cls.predictions.transform.dense.bias', 'cls.seq_relationship.bias', 'cls.seq_relationship.weight', 'cls.predictions.transform.LayerNorm.weight']
- This IS expected if you are initializing BertForSequenceClassification from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).
- This IS NOT expected if you are initializing BertForSequenceClassification from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).
Some weights of BertForSequenceClassification were not initialized from the model checkpoint at bert-base-uncased and are newly initialized: ['classifier.weight', 'classifier.bias']
You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
为了保证到现在为止没有出错，我们放入一个batch看看情况：
outputs = model(**batch)
print(outputs.loss, outputs.logits.shape)
tensor(0.7381, grad_fn=<NllLossBackward>) torch.Size([8, 2])
All 🤗 Transformers models will return the loss when labels are provided, and we also get the logits (two for each input in our batch, so a tensor of size 8 x 2).
接下来还需要一个optimizer和一个learning rate scheduler（learning rate如何调整的机制）：
from transformers import AdamW
optimizer = AdamW(model.parameters(), lr=5e-5)
learning rate默认的调整方法是随着training steps线性降低到0，所以我们需要计算training steps:
from transformers import get_scheduler
num_epochs = 3
# len(train_dataloader) == number of training batches
# 默认情况下取num_epochs为3
num_training_steps = num_epochs * len(train_dataloader)
lr_scheduler = get_scheduler(
"linear",
optimizer=optimizer,
num_warmup_steps=0,
num_training_steps=num_training_steps
)
print(num_training_steps)
1377
The training loop
最后一件事就是保证能在GPU上运行，需要定义device:
import torch
device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
model.to(device)
device
device(type='cuda')
现在可以开始训练啦！为了可视化进度，我们用tqdm定义bar：
from tqdm.auto import tqdm
progress_bar = tqdm(range(num_training_steps))
model.train()
for epoch in range(num_epochs):
for batch in train_dataloader:
batch = {k: v.to(device) for k, v in batch.items()}
outputs = model(**batch)
loss = outputs.loss
loss.backward()
optimizer.step()
lr_scheduler.step()
optimizer.zero_grad()
progress_bar.update(1)
0%| | 0/1377 [00:00<?, ?it/s]
The evaluation loop
上面的训练没有加上evaluation loop，导致我们没有办法看到训练过程中loss，accuracy和F1的变化。现在加上evaluation loop：
from datasets import load_metric
metric= load_metric("glue", "mrpc")
model.eval()
for batch in eval_dataloader:
batch = {k: v.to(device) for k, v in batch.items()}
with torch.no_grad():
outputs = model(**batch)
logits = outputs.logits
predictions = torch.argmax(logits, dim=-1)
# 用add_batch可以累积之前所有batch的结果
metric.add_batch(predictions=predictions, references=batch["labels"])
metric.compute()
Downloading: 0%| | 0.00/1.86k [00:00<?, ?B/s]
{'accuracy': 0.8382352941176471, 'f1': 0.8896321070234114}
Superchange your training loop with 🤗 Accelerate
之前的做法都是在单CPU或单GPU上训练，如果想要在多GPU上进行分布式训练，可以用Hugging Face Accelerate库：
# 导入Accelerator
+ from accelerate import Accelerator
from transformers import AdamW, AutoModelForSequenceClassification, get_scheduler
+ accelerator = Accelerator()
model = AutoModelForSequenceClassification.from_pretrained(checkpoint, num_labels=2)
optimizer = AdamW(model.parameters(), lr=3e-5)
# Accelerator类本身可以知道你的device，不需要自己定义
- device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
- model.to(device)
# 主体调用部分
+ train_dataloader, eval_dataloader, model, optimizer = accelerator.prepare(
+ train_dataloader, eval_dataloader, model, optimizer
+ )
num_epochs = 3
num_training_steps = num_epochs * len(train_dataloader)
lr_scheduler = get_scheduler(
"linear",
optimizer=optimizer,
num_warmup_steps=0,
num_training_steps=num_training_steps
)
progress_bar = tqdm(range(num_training_steps))
model.train()
for epoch in range(num_epochs):
for batch in train_dataloader:
- batch = {k: v.to(device) for k, v in batch.items()}
outputs = model(**batch)
loss = outputs.loss
- loss.backward()
# 将loss replace为accelerator
+ accelerator.backward(loss)
optimizer.step()
lr_scheduler.step()
optimizer.zero_grad()
progress_bar.update(1)
将这段代码放入train.py中：
from accelerate import Accelerator
from transformers import AdamW, AutoModelForSequenceClassification, get_scheduler
accelerator = Accelerator()
model = AutoModelForSequenceClassification.from_pretrained(checkpoint, num_labels=2)
optimizer = AdamW(model.parameters(), lr=3e-5)
train_dl, eval_dl, model, optimizer = accelerator.prepare(
train_dataloader, eval_dataloader, model, optimizer
)
num_epochs = 3
num_training_steps = num_epochs * len(train_dl)
lr_scheduler = get_scheduler(
"linear",
optimizer=optimizer,
num_warmup_steps=0,
num_training_steps=num_training_steps
)
progress_bar = tqdm(range(num_training_steps))
model.train()
for epoch in range(num_epochs):
for batch in train_dl:
outputs = model(**batch)
loss = outputs.loss
accelerator.backward(loss)
optimizer.step()
lr_scheduler.step()
optimizer.zero_grad()
progress_bar.update(1)
然后执行：
!accelerate config
会让你回答一些问题，以此生成config file。然后就可以运行：
!accelerate launch train.py
如果你想在notebook上运行，可以将代码放入函数training_function，然后：
from accelerate import notebook_launcher
notebook_launcher(training_function)