do not convert bf16 to f32 #1055
Merged
+6
−33
Conversation
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Outdated suggestions cannot be applied.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.
Previously, not all ggml ops had good support for BF16 — for example, conv2d/conv3d — so in the code I was forcing BF16 to be converted to F32. However, as more and more models (such as z-image) overflow with FP16 and require BF16, I plan to remove this restriction in the code. For ops that do not support BF16, their weights will need to be converted from BF16 to another type (e.g., FP16 or F32) during weight loading. Developers who need to modify sd.cpp should be aware of this.