Export fp32 to qint4 #3526
Draft
Export fp32 to qint4 #3526
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
Fix problem of group size incosistency Signed-off-by: Grzegorz Kisala/Neural Computing (AIS) /SRPOL/Senior Professional/Samsung Electronics <[email protected]>
Signed-off-by: Grzegorz Kisala/Neural Computing (AIS) /SRPOL/Senior Professional/Samsung Electronics <[email protected]>
Signed-off-by: Grzegorz Kisala/Neural Computing (AIS) /SRPOL/Senior Professional/Samsung Electronics <[email protected]>
Signed-off-by: Grzegorz Kisala/Neural Computing (AIS) /SRPOL/Senior Professional/Samsung Electronics <[email protected]>
Signed-off-by: Grzegorz Kisala/Neural Computing (AIS) /SRPOL/Senior Professional/Samsung Electronics <[email protected]>
Signed-off-by: Grzegorz Kisala/Neural Computing (AIS) /SRPOL/Senior Professional/Samsung Electronics <[email protected]>
**Self evaluation:** 1. Build test: [X]Passed [ ]Failed [ ]Skipped 2. Run test: [X]Passed [ ]Failed [ ]Skipped Signed-off-by: Maciej Nalewaj <[email protected]>
Signed-off-by: Grzegorz Kisala/Neural Computing (AIS) /SRPOL/Senior Professional/Samsung Electronics <[email protected]>
Signed-off-by: Grzegorz Kisala/Neural Computing (AIS) /SRPOL/Senior Professional/Samsung Electronics <[email protected]>
Signed-off-by: Grzegorz Kisala/Neural Computing (AIS) /SRPOL/Senior Professional/Samsung Electronics <[email protected]>
djeong20
requested changes
Oct 20, 2025
| std::vector<uint8_t> quantized_weights_int4; | ||
| std::vector<uint16_t> quantized_scales_int4; | ||
| Int4Utils::quantizeAndRepack(dequantized_weights_q4.data(), N, K, | ||
| Int4Utils::quantizeAndRepack((float *)(file_view + start_from), N, K, |
There was a problem hiding this comment.
FP32 data should be transposed before it is quantized. Otherwise, it would produce an incorrect output.
|
|
||
| std::cout << "New size: " << size << std::endl; | ||
| } else { | ||
| output.write(file_view + start_from, size); |
There was a problem hiding this comment.
In the case of tie word embedding, you need to filter out the LM head case and not save it.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 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.
Dependency of the PR
Commits to be reviewed in this PR
{commit-1}
{commit message}
Self evaluation:
Signed-off-by: {your_name} <{your_email}>
{commit-2}
{commit message}
Self evaluation:
Signed-off-by: {your_name} <{your_email}>
Summary
Signed-off-by: {your_name} <{your_email}>