llava: fixes #12863 v.head.ffn_up.bias byteswap on BE systems #12918
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
Signed-off-by: Aaron Teo <[email protected]>
Signed-off-by: Aaron Teo <[email protected]>
Collaborator
|
If I understand correctly from your original issue, the problem one appears on the last tensor of the file, but not a particular tensor name. So I think this fix is not valid. Imagine in another model, the last tensor is another one, now you will end up with a gguf with 2 broken tensors. |
Collaborator
|
I'm closing this because it's obviously not a valid fix. Even with the same model, you may end up with different tensor order, which may still result in a broken gguf. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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.
This PR fixes the erroneous byteswap for
v.head.ffn_up.biastensor for Big-Endian systems. Please refer to #12863 for more information.I have verified this change on both Little-Endian and Big-Endian systems and it is working as intended. For reference these are the test machines: