Repository navigation
correct-reports: rewrite the fp32 reports already on the Hub - #12
Merged
Merged
Conversation
… they still carry the 8da4w recipe and sizing `python -m pipeline correct-reports <repo> [--apply]` rewrites, in fp32 xnnpack reports only, the recipe wording (group size and int4 codes null, fp32 label and description, named with the toolchain the file was built with) and the sizing (window, table rows, pte estimate), records a `corrections` entry, regenerates the folder's config.json (the app reads its quantization label and fits_phone_budget) and the README, and commits pinned to the revision it read. No .pte is touched. A report is corrected only if its old figures reproduce exactly from the 8da4w sizing with the configured overhead; otherwise it is refused. Dry run on experimentalmachines/SmolLM2-360M-Instruct-ExecuTorch: the 5 xnnpack reports, all reproduced. The corrected .pte estimates are within 1 to 2 percent of the files (1,511,281,131 B estimated against 1,496,334,720 B at 2k); the old ones were 255,300,075 B. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Follow-up to #9. Adds
python -m pipeline correct-reports <repo> [--apply], which rewrites only the fp32 xnnpack reports' recipe wording and sizing, records acorrectionsentry, regenerates the folder'sconfig.json(the app reads its quantization label andfits_phone_budget) and the README, and commits pinned to the revision it read. No.pteis touched. A report is corrected only if its old figures reproduce exactly from the 8da4w sizing, otherwise it is refused.Dry run on
experimentalmachines/SmolLM2-360M-Instruct-ExecuTorch: the 5 xnnpack reports, all reproduced. The corrected.pteestimates are within 1 to 2% of the files (1,511,281,131 B estimated vs 1,496,334,720 B at 2k); the old ones were 255,300,075 B.Tests: 210 passed, 4 skipped.
🤖 Generated with Claude Code
Need help on this PR? Tag
@codesmith-botwith what you need. Autofix is enabled.