MOSS
MOSS copied to clipboard
An open-source tool-augmented conversational language model from Fudan University
想请教一下,在SFT模型训练过程中,MOSS中间模型的验证指标是用的什么,如何选择最优的checkpoint输出,我这边尝试使用了BLEU等指标,效果并不好,不能反应模型的实际效果
LMSYS org was rolling out some serious back to back battle between LLMs, I would like you see MOSS's debut there 👍
ERROR: Could not find a version that satisfies the requirement triton (from versions: none) ERROR: No matching distribution found for triton
环境:ubuntu20.04 torch 1.10.1 transformers 4.27.1 code: import os import torch from huggingface_hub import snapshot_download from transformers import AutoConfig, AutoTokenizer, AutoModelForCausalLM from accelerate import init_empty_weights, load_checkpoint_and_dispatch os.environ['CUDA_VISIBLE_DEVICES'] = "0,1" model_path =...
response之后需要添加 ``
可以进到顺利打开webui。 输入文字后报错。 RuntimeError: CUDA error: device-side assert triggered CUDA kernel errors might be asynchronously reported at some other API call,so the stacktrace below might be incorrect. For debugging consider passing...

第一条数据就错了,只有直角三角形才能用勾股定理。这模型都被整数据的带沟里去了 moss/SFT_data/conversations/conversation_with_plugins/calculator/calculator_plugin_conversation_1.json "chat": { "turn_1": { "Human": ": 一个三角形的其中两边分别是16,12,第三条边长度是多少?\n", "Inner Thoughts": ": 这是一个基本的几何问题,需要使用勾股定理来求解。设第三条边为x,则有16^2=12^2+x^2,可以求解得到x=8√5\n", "Commands": ": None\n", "Tool Responses": ":\nNone\n\n", "MOSS": ": 这个三角形的第三条边长度是8√5。\n" },
已经 installed transformers from source , 并且 installed sentencepiece,运行 run.sh 微调时还是会报这个错误