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

Add DeepSeek V2.5 Example #171

Merged
merged 3 commits into from
Sep 13, 2024
Merged

Add DeepSeek V2.5 Example #171

merged 3 commits into from
Sep 13, 2024

Conversation

dsikka
Copy link
Collaborator

@dsikka dsikka commented Sep 12, 2024

SUMMARY:

  • Add an example for quantizing deepseek v2.5 and running on vLLM

Copy link

👋 Hi! Thank you for contributing to llm-compressor. Please add the ready label when the PR is ready for review.

Copy link
Member

@mgoin mgoin left a comment

Choose a reason for hiding this comment

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

LGTM only nits for future consideration

MODEL_ID,
reserve_for_hessians=True,
num_gpus=2,
torch_dtype=torch.bfloat16,
Copy link
Member

Choose a reason for hiding this comment

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

In the future we should rely on auto dtype, but it seems like deepseek v2.5 is bfloat16 so this is okay

ignore: [lm_head, "re:.*mlp.gate$"]
config_groups:
group_0:
weights: {num_bits: 4, type: int, symmetric: true, strategy: channel, dynamic: false}
Copy link
Member

Choose a reason for hiding this comment

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

We may want to use grouped and actorder for the final example based on how the accuracy comes out, but this is good for now

@dsikka dsikka merged commit 4fe45de into main Sep 13, 2024
5 of 7 checks passed
@dsikka dsikka deleted the add_deepseek_moe_example branch September 13, 2024 15:31
markmc pushed a commit to markmc/llm-compressor that referenced this pull request Nov 13, 2024
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