-
Notifications
You must be signed in to change notification settings - Fork 1
/
3-transformer-tune.py
70 lines (54 loc) · 1.73 KB
/
3-transformer-tune.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
import torch
from transformers import GPT2Tokenizer, GPT2LMHeadModel, Trainer, TrainingArguments, DataCollatorForLanguageModeling
from datasets import load_from_disk
dataset = load_from_disk('./shakespeare_sentences')
tokenizer = GPT2Tokenizer.from_pretrained("gpt2")
tokenizer.pad_token = tokenizer.eos_token
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
torch.manual_seed(12345)
if torch.cuda.is_available():
torch.cuda.manual_seed(12345)
model = GPT2LMHeadModel.from_pretrained("gpt2")
model.to(device)
def tokenize_function(examples):
return tokenizer(examples['sentence'], truncation=True, max_length=128)
tokenized_dataset = dataset.map(tokenize_function, batched=True, remove_columns=["sentence"])
data_collator = DataCollatorForLanguageModeling(
tokenizer=tokenizer, mlm=False,
)
training_args = TrainingArguments(
output_dir='./results',
overwrite_output_dir=True,
num_train_epochs=1,
per_device_train_batch_size=8,
save_steps=500,
save_total_limit=2,
logging_steps=100,
prediction_loss_only=True,
learning_rate=5e-4,
no_cuda=not torch.cuda.is_available(),
)
trainer = Trainer(
model=model,
args=training_args,
train_dataset=tokenized_dataset['test'],
data_collator=data_collator,
)
trainer.train()
trainer.save_model('./model_1')
prompt = "To be, or not to be"
input_ids = tokenizer.encode(prompt, return_tensors='pt').to(device)
model.eval()
output = model.generate(
input_ids,
max_length=128,
do_sample=True,
top_k=50,
top_p=0.95,
temperature=0.7,
num_return_sequences=1,
no_repeat_ngram_size=2,
early_stopping=True,
)
generated_text = tokenizer.decode(output[0], skip_special_tokens=True)
print(generated_text)