Skip to content

Conversation

zhyncs
Copy link
Member

@zhyncs zhyncs commented Jan 2, 2025

Motivation

  • EAGLE-2
  • DeepSeek V3 36 tokens/s

Modifications

Checklist

  • Format your code according to the Contributor Guide.
  • Add unit tests as outlined in the Contributor Guide.
  • Update documentation as needed, including docstrings or example tutorials.

@zhyncs zhyncs self-assigned this Jan 2, 2025
@zhyncs
Copy link
Member Author

zhyncs commented Jan 2, 2025

We will wait for the EAGLE feature to stabilize before releasing post4 and add automated unit tests, benchmarks, and evaluations.

@zhyncs
Copy link
Member Author

zhyncs commented Jan 5, 2025

I will first release a new version, and these todos are left for post5.

@zhyncs zhyncs merged commit 2f0d386 into main Jan 5, 2025
17 checks passed
@zhyncs zhyncs deleted the zhyncs/v0.4.1.post4 branch January 5, 2025 17:29
@aisensiy
Copy link

aisensiy commented Jan 6, 2025

Right now deploy using two H800 GPUs (8 per node) to achieve an inference speed of 18.01 t/s.

timethink pushed a commit to timethink/sglang that referenced this pull request Mar 9, 2025
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.

2 participants