chensongcan
chensongcan
could you provide the dataset？
@Sanqiang Could you please provide the file for dataset？ /zfs1/hdaqing/saz31/dataset/
@DhairyaPatel7 have your solved your problem？
SO am I，could you tell me whether your problem has been solved or not?
我的数据加载是先处理完cache存起来加载的，目前已经是进入训练阶段，根据错误提示，应该是卡在了all_gather阶段 补充一下： 一样的环境跑其他的模型，如qwen1.5非moe模型没有问题，可以正常训练
Restricted text generation, I can think of the direction of landing is to make a sentence, I do not know what other direction of landing. Could you share some of...
使用预训练脚本，通过去掉overwrite_output_dir参数，加载保存在output的checkpoint-100文件夹，出现RuntimeError: Error(s) in loading state_dict for PeftModelForCausalLM:
请问目前公开的代码预训练的时候加上bias项了吗
这个项目是基于llama从0开始训练是吗