Repository navigation
Conversation
This was referenced Oct 29, 2026
jiapingW
force-pushed
the
split/config-03-components
branch
from
October 29, 2026 09:20
a4de98e to
be68499
Compare
jiapingW
changed the title
This was referenced Oct 30, 2026
jiapingW
changed the title
jiapingW
changed the base branch from
split/config-02-options
to
split/config-weights
November 1, 2026 03:10
This was referenced Nov 1, 2026
jiapingW
force-pushed
the
split/config-weights
branch
from
November 1, 2026 06:11
3c46ae2 to
eaa788e
Compare
jiapingW
force-pushed
the
split/config-03-components
branch
from
November 1, 2026 06:11
e80e815 to
e9d1723
Compare
This branch has not been 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Move models into modeling/draft/models and update imports and mock targets. Rename llama3_eagle.py to models/eagle3.py; place convolution, attention backends and ring parallelism in their own layer packages. Move the remaining tensor-parallel embedding, linear and LM head primitives from specforge/layers/ into models/layers/parallel/, and update their tests and ownership rules. Remove the old import paths.
Component extraction follows in a separate PR.
Depends on #1059.
Validation: 637 CPU-suite tests completed, 48 skipped. Source files are moved without logic changes. Tensor-parallel components and their GPU test modules import successfully; multi-GPU execution was not exercised.
Stack: #1042 → #1043 → #1058 → #1059 → #1044 → #1060 → #1061 → #1052 → #1045 → #1046.
Independent CPU CI: #1053.