[GH-ISSUE #155] 微调完的模型,如何跟基础模型合并? #133

Open
opened 2026-10-26 21:03:58 +00:00 by bsakaguchi · 1 comment
Owner

Originally created by @libingbingd on GitHub (Apr 16, 2024).
Original GitHub issue: https://github.com/deepseek-ai/DeepSeek-Coder/issues/155

您好,我微调完发现没有大的模型权重文件,只有一些配置文件,如何跟原模型合并呢?
微调后文件列表

Originally created by @libingbingd on GitHub (Apr 16, 2024). Original GitHub issue: https://github.com/deepseek-ai/DeepSeek-Coder/issues/155 您好,我微调完发现没有大的模型权重文件,只有一些配置文件,如何跟原模型合并呢? <img width="178" alt="微调后文件列表" src="https://github.com/deepseek-ai/DeepSeek-Coder/assets/147381086/f02defdc-d5b5-431e-8bf6-19e4cdd662a8">
Author
Owner

@DejianYang commented on GitHub (Apr 18, 2024):

您好,我微调完发现没有大的模型权重文件,只有一些配置文件,如何跟原模型合并呢? 微调后文件列表

检查下是不是你的训练步数太少 没有save checkpoitn

<!-- gh-comment-id:2063068103 --> @DejianYang commented on GitHub (Apr 18, 2024): > 您好,我微调完发现没有大的模型权重文件,只有一些配置文件,如何跟原模型合并呢? <img alt="微调后文件列表" width="178" src="https://private-user-images.githubusercontent.com/147381086/322758699-f02defdc-d5b5-431e-8bf6-19e4cdd662a8.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmF3LmdpdGh1YnVzZXJjb250ZW50LmNvbSIsImtleSI6ImtleTUiLCJleHAiOjE3MTM0MjA1NTMsIm5iZiI6MTcxMzQyMDI1MywicGF0aCI6Ii8xNDczODEwODYvMzIyNzU4Njk5LWYwMmRlZmRjLWQ1YjUtNDMxZS04YmY2LTE5ZTRjZGQ2NjJhOC5wbmc_WC1BbXotQWxnb3JpdGhtPUFXUzQtSE1BQy1TSEEyNTYmWC1BbXotQ3JlZGVudGlhbD1BS0lBVkNPRFlMU0E1M1BRSzRaQSUyRjIwMjQwNDE4JTJGdXMtZWFzdC0xJTJGczMlMkZhd3M0X3JlcXVlc3QmWC1BbXotRGF0ZT0yMDI0MDQxOFQwNjA0MTNaJlgtQW16LUV4cGlyZXM9MzAwJlgtQW16LVNpZ25hdHVyZT1iMGY3ZWIxNGZiODcxODJjNTc2ZDA2Zjg5OWZiMmM0MzIzMThmZGJkNTQ5YzY1MWJhMzJiNWM1MGFkYTQ5NDU1JlgtQW16LVNpZ25lZEhlYWRlcnM9aG9zdCZhY3Rvcl9pZD0wJmtleV9pZD0wJnJlcG9faWQ9MCJ9.QJmoxIXELqjIpr9QaOKTvlRlwvq78pRPs_-zSxFB__g"> 检查下是不是你的训练步数太少 没有save checkpoitn
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
deepseek-ai/DeepSeek-Coder#133
No description provided.