Fix DeepSpeed 0.17 compatibility (bf16 config + nested autocast)#1063
Merged
Conversation
DeepSpeed 0.17.x validates configs with pydantic (extra='forbid') and asserts against torch.autocast nested outside the engine. - Remove invalid `loss_scale_window` from the bf16 ds_config block; it is only valid for fp16 (DeepSpeedBF16Config rejects it). Older DeepSpeed silently ignored it. - Disable the forward-pass torch.autocast wrapper when using DeepSpeed, since the engine handles mixed precision via its bf16/fp16 config. Matches the existing gating in the backward pass and eval loop. Co-Authored-By: Claude Opus 4.8 (1M context) <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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
DeepSpeed 0.17.x validates configs with pydantic (extra='forbid') and asserts against torch.autocast nested outside the engine.
loss_scale_windowfrom the bf16 ds_config block; it is only valid for fp16 (DeepSpeedBF16Config rejects it). Older DeepSpeed silently ignored it.closes #945