求助贴--大模型04期使用xtuner微调开源大模型 #392
Labels
No Label
bug
duplicate
enhancement
help wanted
invalid
question
wontfix
No Milestone
No project
No Assignees
2 Participants
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: HswOAuth/llm_course#392
Loading…
Reference in New Issue
Block a user
No description provided.
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
按照教材,进行单机多卡微调训练操作,模型为Baichuan-7B,数据为:timdettmers/openassistant-guanaco ,配置文件做了简单修改,主要是模型和数据的目录修改,详见图一。执行命令参考教程,NPROC_PER_NODE=4 xtuner train /code/llama2_7b_chat_qlora_alpaca_e3_copy.py --deepspeed deepspeed_zero3
报错信息详见图二,11/20 16:46:21 - mmengine - INFO - xtuner_dataset_timeout = 0:30:00
Repo card metadata block was not found. Setting CardData to empty.
Generating train split: 9846 examples [00:00, 49692.76 examples/s]
Generating test split: 518 examples [00:00, 26114.20 examples/s]
附图一:

附图二:

看上去好像还是微调数据库的问题,不过数据都是正常下载的,不应该有问题。详见附图三。
附图一

附图二

附图三

看训练模版,应该是要使用alpaca数据集格式;
如果想用百川模型,需要使用百川的训练模板
可以使用这个模板:
xtuner copy-cfg baichuan_7b_qlora_oasst1_e3 .
教程不就是这么写的吗?
这里写的是如何下载数据和模型,教程了使用的是llama作为训练模版。
想要使用baichuan模型请使用这个训练模板,命令是:
xtuner copy-cfg baichuan_7b_qlora_oasst1_e3 .