DeepSpeed
fix: Validate fp16.loss_scale is finite and non-negative
#7889
Merged

fix: Validate fp16.loss_scale is finite and non-negative #7889

nathon-lee
nathon-lee nathon-lee requested a review from tjruwase tjruwase 35 days ago
nathon-lee nathon-lee requested a review from tohtana tohtana 35 days ago
chatgpt-codex-connector
chatgpt-codex-connector commented on 2026-03-06
nathon-lee nathon-lee changed the title Validate fp16.loss_scale is finite and non-negative fix: Validate fp16.loss_scale is finite and non-negative 35 days ago
PKUWZP PKUWZP requested a review from PKUWZP PKUWZP 34 days ago
PKUWZP
PKUWZP requested changes on 2026-03-06
nathon-lee nathon-lee force pushed from f0059a79 to 5c7b12e7 34 days ago
nathon-lee nathon-lee requested a review from loadams loadams 34 days ago
nathon-lee nathon-lee requested a review from hwchen2017 hwchen2017 34 days ago
nathon-lee nathon-lee requested a review from GuanhuaWang GuanhuaWang 34 days ago
nathon-lee nathon-lee force pushed from 5c7b12e7 to f0059a79 34 days ago
nathon-lee nathon-lee force pushed from f0059a79 to 3ead20d9 34 days ago
nathon-lee fix: Validate fp16.loss_scale is finite and non-negative
225ab4e5
nathon-lee nathon-lee force pushed from 3ead20d9 to 225ab4e5 34 days ago
nathon-lee fix: validate fp16.loss_scale before coercion
f2fc3092
nathon-lee nathon-lee force pushed from 9ad6000c to f2fc3092 34 days ago
nathon-lee nathon-lee force pushed from 403cd4c6 to 39b2b3a8 32 days ago
nathon-lee Refactor loss scale validation in precision_config.py
0ac46845
nathon-lee chore: DCO signoff
322707ab
nathon-lee nathon-lee force pushed from 39b2b3a8 to 322707ab 32 days ago
nathon-lee test: ensure invalid fp16.loss_scale types raise ValidationError
a8cb9dca
tohtana Merge branch 'master' into fix_issue_7852
dcd1e056
tohtana
tohtana approved these changes on 2026-03-10
nathon-lee Fix: remove trailing whitespace, yapf format, update license header (…
e59f7b8f
nathon-lee Merge branch 'fix_issue_7852' of github.com:nathon-lee/DeepSpeed_woo …
a29c6ea8
nathon-lee Add license header to precision config loss scale unit test
85f38218
tohtana Merge branch 'master' into fix_issue_7852
b7bb966a
tohtana tohtana enabled auto-merge (squash) 28 days ago
tohtana Merge branch 'master' into fix_issue_7852
63744bd4
tohtana
PKUWZP PKUWZP requested a review from PKUWZP PKUWZP 28 days ago
PKUWZP
PKUWZP approved these changes on 2026-03-13
tohtana Merge branch 'master' into fix_issue_7852
0708001a
tohtana tohtana merged 63eeb114 into master 28 days ago

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone