ymcui / Chinese-LLaMA-Alpaca-2

中文LLaMA-2 & Alpaca-2大模型二期项目 + 64K超长上下文模型 (Chinese LLaMA-2 & Alpaca-2 LLMs with 64K long context models)
Apache License 2.0
7k stars 570 forks source link

预训练完成后模型的使用 #548

Closed ymourenya closed 2 months ago

ymourenya commented 3 months ago

提交前必须检查以下项目

问题类型

模型训练与精调

基础模型

Chinese-LLaMA-2 (7B/13B)

操作系统

Windows

详细描述问题

# 请在此处粘贴运行代码(请粘贴在本代码块里)

依赖情况(代码类问题务必提供)

# 请在此处粘贴依赖情况(请粘贴在本代码块里)

运行日志或截图

你好,我在预训练的使用使用了参数modules_to_save="embed_tokens,lm_head" ,请问训练好的参数会保存到图中的模型吗,我可以直接删除原模型使用图中的模型+lora模块吗 capture_20240325140115684

ymourenya commented 3 months ago

请问这两个模型分别是什么,第一个25G多,第二个7G多

iMountTai commented 3 months ago

global_step20下面是训练中状态文件,是deepspeed的权重文件。训练好的参数应该是保存在了checkpoint-20根目录下,名称为adapter_model.binadapter_model.safetensors

github-actions[bot] commented 2 months ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your consideration.

github-actions[bot] commented 2 months ago

Closing the issue, since no updates observed. Feel free to re-open if you need any further assistance.