Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

LLM: Enable Speculative on Fastchat #10909

Merged
merged 8 commits into from
May 6, 2024

Conversation

hzjane
Copy link
Contributor

@hzjane hzjane commented Apr 29, 2024

Description

Enable Speculative on Fastchat
Llama-7b speculative will use near 16.5GB(fp16+int4) memory on Self-Speculative case,So just can run on max GPU.

1. Why the change?

2. User API changes

3. Summary of the change

4. How to test?

  • Application test on spr14
  • Application test on max1100

5. New dependencies

  • New Python dependencies
    - Dependency1
    - Dependency2
    - ...
  • New Java/Scala dependencies and their license
    - Dependency1 and license1
    - Dependency2 and license2
    - ...

@hzjane hzjane requested review from glorysdj and gc-fu April 29, 2024 07:40
@hzjane hzjane requested a review from glorysdj April 29, 2024 08:12
@glorysdj
Copy link
Contributor

@hzjane hzjane marked this pull request as ready for review April 29, 2024 08:43

```bash
# Available low_bit format including bf16 on CPU, fp16 on MAX GPU.
python3 -m ipex_llm.serving.fastchat.ipex_llm_worker --model-path lmsys/vicuna-7b-v1.5 --low-bit "bf16" --trust-remote-code --device "cpu" --speculative
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

also add gpu example

Copy link
Contributor

@gc-fu gc-fu left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The code changes LGTM

@hzjane hzjane merged commit 0e0bd30 into intel-analytics:main May 6, 2024
18 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

3 participants