Refactor: flatten models/ to one directory per model build - #886
Open
zhangqi-chen wants to merge 1 commit into
Open
Refactor: flatten models/ to one directory per model build#886zhangqi-chen wants to merge 1 commit into
zhangqi-chen wants to merge 1 commit into
Conversation
|
Important Review skippedToo many files! This PR contains 184 files, which is 84 over the limit of 100. To get a review, narrow the scope: Upgrade to a paid plan to raise the limit. This review couldn't start because sufficient usage credits or metered capacity aren't available. Add credits or update usage-based reviews in the billing tab, then retry. ⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Plus Run ID: ⛔ Files ignored due to path filters (1)
📒 Files selected for processing (184)
You can disable this status message by setting the Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
The models/<family>/<variant>/ nesting held no family-level shared code, so the intermediate directories were pure namespacing. Flatten them to <family>_<version>[_<quant>], which also settles the hyphen-vs-underscore inconsistency between v4-flash and v3_2. - qwen3/14b, qwen3/32b, deepseek/v3_2 become qwen3_14b, qwen3_32b, deepseek_v3_2. - deepseek/v4-flash becomes deepseek_v4_flash_w8a8int and deepseek/v4-pro becomes deepseek_v4_pro_w4a8mx. The suffix names the deployment quantization with the CANN convert_model.py --quant_type vocabulary, separators dropped. Flash matches its Atlas A3 deployment; Pro names its A5 Hybrid MXFP8-MXFP4 target while its kernels still consume INT8 re-quantized from the MXFP4 grid. - Drop the now-redundant model-name prefix from filenames under deepseek_v3_2, qwen3_32b, and the qwen3_14b drafts. - contract.py derived the models root as _KERNEL_DIR.parent.parent.parent, one level too many for the flatter tree; use .parent, which matches the variable name. config.py and constants.py exist only under models/, so the narrower prefix keeps the shadowed-module check equivalent. - Retarget the CI change selectors, the contract registry variant path, the contract test's Path-segment literals, and the docs.
zhangqi-chen
force-pushed
the
refactor/flatten-models
branch
from
August 3, 2026 06:22
f77fec9 to
9b59d7f
Compare
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
1 participant
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.
The models/// nesting held no family-level shared code,
so the intermediate directories were pure namespacing. Flatten them to
[], which also settles the hyphen-vs-underscore
inconsistency between v4-flash and v3_2.
deepseek_v3_2.
becomes deepseek_v4_pro_w4a8mx. The suffix names the deployment
quantization with the CANN convert_model.py --quant_type vocabulary,
separators dropped. Flash matches its Atlas A3 deployment; Pro names
its A5 Hybrid MXFP8-MXFP4 target while its kernels still consume INT8
re-quantized from the MXFP4 grid.
deepseek_v3_2, qwen3_32b, and the qwen3_14b drafts.
one level too many for the flatter tree; use .parent, which matches the
variable name. config.py and constants.py exist only under models/, so
the narrower prefix keeps the shadowed-module check equivalent.
the contract test's Path-segment literals, and the docs.