Skip to content

Feat/tensor override #1180

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

Open
wants to merge 2 commits into
base: master
Choose a base branch
from
Open

Conversation

dpmm99
Copy link
Contributor

@dpmm99 dpmm99 commented May 2, 2025

Based on the April_2025 branch. Should I rebase it?

Example usage:

var modelParams = new ModelParams(path)
{
    TensorBufferOverrides = [new Abstractions.TensorBufferOverride(".*ffn.*", "CPU")],
    GpuLayerCount = -1
};

would make all tensors with 'ffn' in their names offload to CPU, with everything else on GPU.

@martindevans
Copy link
Member

martindevans commented May 2, 2025

Should I rebase it?

yes please, that'll remove a lot of noise from this PR and make it easier to review.

@dpmm99 dpmm99 force-pushed the feat/tensor-override branch from 3dd6110 to 8dc45de Compare May 2, 2025 20:19
@dpmm99
Copy link
Contributor Author

dpmm99 commented May 2, 2025

Should I rebase it?

yes please, that'll remove a lot of noise from this PR and make it easier to review.

Okay, done. But now it's going to show a conflict when you merge the Apr_2025 branch because they both "added" tensor_buft_overrides to LLamaModelParams. :P

@martindevans
Copy link
Member

CI is failing due to ot being detected as a spelling error, just add a suppression for that.

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