Repository navigation
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/24946
Note: Links to docs will display an error until the docs builds have been completed. ❌ 133 New Failures, 8 Unrelated Failures, 10 Unclassified FailuresAs of commit c399e30 with merge base c7f7931 ( NEW FAILURES - The following jobs have failed:
UNCLASSIFIED FAILURES - DrCI could not classify the following jobs because the workflow did not run on the merge base. The failures may be pre-existing on trunk or introduced by this PR:
FLAKY - The following jobs failed but were likely due to flakiness present on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
Bot
added
the
CLA Signed
This was referenced Nov 12, 2026
transformers no longer re-exports accelerate's init_empty_weights, so export_hf_llm failed to import. Build GGUF-loaded models with a local context that puts parameters on the meta device and leaves buffers on the CPU, as init_empty_weights did; a plain meta device would also empty computed buffers such as RoPE frequencies. Pass dtype rather than the deprecated torch_dtype. Newer StaticLayer.lazy_initialization also takes the value states, so the FP16 KV cache passes them where the signature asks. The tests cover both versions: the GGUF mapping test uses eager attention, since newer sdpa rejects Gemma 2's softcap, and the FP16 sliding-window rejection test skips where transformers keeps sliding layers in StaticLayer and windows them through the mask. Tested with transformers 5.0.0rc1 and 5.22.2.
metascroy
force-pushed
the
offgraph-kv-mg-ops
branch
from
November 14, 2026 00:18
9bc8b60 to
ddb4ff2
Compare
metascroy
force-pushed
the
offgraph-kv-transformers-compat
branch
from
November 14, 2026 00:18
ea18142 to
c399e30
Compare
This branch had an error being deployed
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
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.
Stack
Off-graph KV cache for the native ggml engine, through serving Muse Glimmer on it. Review bottom-up; each PR builds and passes its tests on its own.
Summary
export_hf_llm.pyfailed to import on newer transformers, which no longer re-exports accelerate'sinit_empty_weights.init_empty_weightsdid. A plain meta device would also empty computed buffers such as RoPE frequencies.dtypereplaces the deprecatedtorch_dtype.StaticLayer.lazy_initializationalso takes the value states, so the FP16 KV cache passes them where the signature asks.StaticLayerand windows them through the mask.Test plan
test_export_hf_llm.pypasses with transformers 5.0.0rc1 (the examples pin) and 5.22.2.Authored with Claude Code.