README.md

metadata

language:
  - cn
  - en
license: mit
datasets:
  - TigerResearch/tigerbot-zhihu-zh-10k
pipeline_tag: text-generation

中文文本生成

1 Usage

1.1 Initalization

!pip install transformers[torch]

from transformers import GPT2Tokenizer, GPT2LMHeadModel
import torch
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

tokenizer = GPT2Tokenizer.from_pretrained('Hollway/gpt2_finetune')
model = GPT2LMHeadModel.from_pretrained('Hollway/gpt2_finetune').to(device)

1.2 Inference

def generate(text):
    inputs = tokenizer(text, return_tensors="pt").to(model.device)
    with torch.no_grad():
        tokens = model.generate(
            **inputs,
            max_new_tokens=256,
            do_sample=True,
            temperature=0.7,
            top_p=0.9,
            repetition_penalty=1.05,
            pad_token_id=tokenizer.pad_token_id,
        )
    return tokenizer.decode(tokens[0], skip_special_tokens=True)

generate("只因你")