kvcache-ai / kvcache-ai/ktransformers

how to convert Kimi-K2.5 base model is in INT4 format to BF16 before SFT

Open
#1,829 5 comments 0 reactions 2 assignees Claimed by @JimmyPeilinLi View on GitHub
Dominant language
Python
Stars
19.5k
Forks
1.6k
Avg merge
19h 32m
Merged PRs (30d)
27

Description

### Reminder

- [x] I have read the above rules and searched the existing issues.

### System Info

Hello, I wan to sft or lora K2.5, In https://github.com/kvcache-ai/ktransformers/blob/main/doc/en/SFT_Installation_Guide_KimiK2.5.md, it says:
```
1.2 Convert INT4 → BF16
Kimi-K2.5 base model is in INT4 format, convert it to BF16 before SFT.
```
But how to convert from INT4 to BF16, any scripts?

### Reproduction

```text
Put your message here.
```

### Others

_No response_

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.