爱心源码-一个基于最大的开源GPT2中文大模型的写作机器人源码

#from transformers import GPT2Model; \ #from deepspeed.runtime.zero.stage3 import estimate_zero3_model_states_mem_needs_all_live; \ #model = GPT2Model.from_pretrained("IDEA-CCNL/Wenzhong2.0-GPT2-3.5B-chinese"); \ #estimate_zero3_model_states_mem_needs_all_live(model, num_gpus_per_node=8, num_nodes=1) from transformers import GPT2Tokenizer,GPT2LMHeadModel hf_model_path = 'IDEA-CCNL/Wenzhong2.0-GPT2-3.5B-chinese' tokenizer = GPT2Tokenizer.from_pretrained(hf_model_path) model = GPT2LMHeadModel.from_pretrained(hf_model_path) question = "解放军是威武文明的" inputs = tokenizer(question,return_tensors='pt') generation_output = model.generate(**inputs, return_dict_in_generate=True, output_scores=True, max_length=150, # max_new_tokens=80, do_sample=True, top_p = 0.6, # num_beams=5, eos_token_id=50256, pad_token_id=0, num_return_sequences = 5) for idx,sentence in enumerate(generation_output.sequences): print('next sentence %d:\n'%idx, tokenizer.decode(sentence).split('<|endoftext|>')[0]) print('*'*40)