TP: fix Phi3, Bert, Plamo2/3, ChatGLM - #25536
Merged
Merged
Conversation
ggerganov
approved these changes
Jul 15, 2026
JohannesGaessler
force-pushed
the
tp-phi3
branch
from
July 16, 2026 12:09
ecad51a to
f0a0baf
Compare
Member
|
@JohannesGaessler Good to merge? |
Contributor
Author
|
From my end yes, I just wanted to rebase on top of the new virtual GPU support. |
satindergrewal
pushed a commit
to satindergrewal/llama.cpp
that referenced
this pull request
Aug 12, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Alternative to #24936 .
Some models like Phi 3 have "ffn_up" tensors that are actually a fused up and gate tensor. This in principle causes a failure in
test-llama-archsbut only if executed on 2+ GPUs because the segmenting happens to be accidentally correct for 1 GPU; for this reason we did not detect the failure in our CI. Since there are multiple models that have this issue and it's easy to forget the in my opinion correct way to fix this is to check the "ffn_up" tensor shape and infer whether or not it is a fused up + gate from that.The change to
test-llama-arch.cppis simply so that I can more easily inspect the seeds being used with a debugger.Requirements