leizhu1989

@steven8274 thank you for your reply.

> I think 【train_prompts.py】 is the first step to train SFT, 【train_reward_models.py】 is the second step to train RM, 【train_dummy.py】 uses PPO training, initial_model uses the model of the first...

> Looking at the paper, the first and second steps use prompt data, and the last step does not seem to require prompt data. I'm not sure either. In addition,...

> As for the model training, I am also exploring it, and there is a lack of data at the moment. ok,my qq:805650606

command：iopaint start --model-dir=D:\project\OCR\PaddleOCR\model --model=lama --device=cpu --port=8080 os: windows11 conda env: cuda118+torch2.1.2 result： 2024-04-30 14:30:11.744 | INFO | iopaint.runtime:setup_model_dir:82 - Model directory: D:\project\OCR\PaddleOCR\model - Platform: Windows-10-10.0.22631-SP0 - Python version: 3.11.7 -...

I change to linux , then normal

好的，谢谢回复，是可以搜索任意自己有关语料的吧，很期待

非常期待

@ 大佬，支持自定义知识库QA问答搜索的话，可以高职一下吗，谢谢了

推理代码： from transformers import AutoModelForCausalLM, AutoTokenizer tokenizer = AutoTokenizer.from_pretrained("baichuan-inc/baichuan-7B", trust_remote_code=True) model = AutoModelForCausalLM.from_pretrained("baichuan-inc/baichuan-7B", device_map="auto", trust_remote_code=True) inputs = tokenizer('登鹳雀楼->王之涣\n夜雨寄北->', return_tensors='pt') inputs = inputs.to('cuda:0') pred = model.generate(**inputs, max_new_tokens=64,repetition_penalty=1.1) print(tokenizer.decode(pred.cpu()[0], skip_special_tokens=True))