Port 24.04 changes into main #851
Merged
Conversation
This file contains 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
**Changes:**
- New option flags `--pax-split-qkv` and
`--te-qkv-layout={qkv_packed,kv_packed}` for identifying QKV layouts for
both the Pax source and TE target. Default behavior without new flags
remains unchanged.
- New option flag `--use-gated-activations` to support gated GELU in
LLaMa2.
- Bias terms are omitted from the checkpoint map when `--weight-only` is
used.
**Notes:**
- Changes are verified with a 70b LLaMa2 checkpoint previously converted
with @mingxu1067's hard-coded LLaMa2-only converter script.
```
python converter/main.py \
--input-path=<...> \
--output-path=<...> \
--fw=pax --direction=fw2te \
--num-of-layer=80 \
--num-of-head=64 \
--head-dim=128 \
--mlp-intermediate-dim=28672 \
--kernel-chunk-size=512 \
--weight-only \
--use-gated-activations \
--pax-split-qkv \
--te-qkv-layout=kv_packed
```
---------
Signed-off-by: Alp Dener <[email protected]>
Co-authored-by: ashors1 <[email protected]>
terrykong
requested changes
May 29, 2024
terrykong
previously approved these changes
May 29, 2024
terrykong
approved these changes
Jun 7, 2024
|
LGTM. I cancelled the most recent CI since the most recent change was just to markdown and previous CI run showed no unexpected errors. |
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Outdated suggestions cannot be applied.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.
No description provided.