Llama4scout logit checking #475
Open
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: George Polovets <gpolovets@gmail.com>
…y are downloaded remotely. Also adding llama4_attention.py because I previously forgot. Signed-off-by: George Polovets <gpolovets@gmail.com>
…g. Also fixed RMSNorm to divide by the std instead of multiply
…ply router scores before inputing to experts. Still getting gibberish output though. Signed-off-by: George Polovets <gpolovets@gmail.com>
… to reshape the correct HF axis (it is the inverse of the QKV shapes). Now not producing gibberish but need to confirm that decode and MMLU quality is maintained. Signed-off-by: George Polovets <gpolovets@gmail.com>
…Llama3, and adding printing of the model architecture.
…rings from latest mainline changes and brought back forking into model_loader to support random weights. Accuracy/perf is roughly the same as before the rebase. (code also only works with recent vLLM pinned checkpoint)
…tting in half. Now the logits match with pytorch implementation.
…ed by both systems are similar but usually not exact. Signed-off-by: George Polovets <gpolovets@gmail.com>
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Outdated suggestions cannot be applied.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.
Description
Start with a short description of what the PR does and how this is a change from
the past.
The rest of the description includes relevant details and context, examples:
If the change fixes a bug or a Github issue, please include a link, e.g.,:
FIXES: b/123456
FIXES: #123456
Tests
Please describe how you tested this change, and include any instructions and/or
commands to reproduce.
Checklist
Before submitting this PR, please make sure: