8000 vulkan: mark IM2COL as supporting non-contig by jeffbolznv · Pull Request #13783 · ggml-org/llama.cpp · GitHub
[go: up one dir, main page]

Skip to content

vulkan: mark IM2COL as supporting non-contig #13783

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

Merged
merged 1 commit into from
May 26, 2025

Conversation

jeffbolznv
Copy link
Collaborator

Fixes #13778 and #13597.

The src0 tensor is not actually used, so it can be non-contiguous.

@jeffbolznv jeffbolznv requested a review from 0cc4m May 25, 2025 20:23
@github-actions github-actions bot added Vulkan Issues specific to the Vulkan backend ggml changes relating to the ggml tensor library for machine learning labels May 25, 2025
Copy link
Collaborator
@0cc4m 0cc4m left a comment

Choose a reason for hiding this comment

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

That's odd, so it's just taking the shape from src0?

Change looks good to me.

@0cc4m 0cc4m merged commit fef693d into ggml-org:master May 26, 2025
46 checks passed
@jeffbolznv
Copy link
Collaborator Author

I was a bit surprised too, but that seems to be the case in the CPU backend as well. I suspect it's so if you had a convolution operation you could pass the kernel tensor in to indicate the shape, rather than having to extract that shape into separate parameters.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
ggml changes relating to the ggml tensor library for machine learning Vulkan Issues specific to the Vulkan backend
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Eval bug: Mistral Small Multiomodal fails when used with the Vulkan backend
2 participants
0