NVIDIA / NVIDIA/Megatron-LM

Support for gated attention in MLA

Open
#4,837 2 comments 0 reactions 1 assignee Claimed by @janEbert View on GitHub
community-request enhancement waiting-on-maintainers
Dominant language
Python
Stars
17.9k
Forks
4.5k
Avg merge
4d 6h
Merged PRs (30d)
271

Description

**Is your feature request related to a problem? Please describe.**
I noticed that in `megatron/core/transformers/transformers_config.py`, there is a comment on line 2428 stating:
```python
# Output gate is not supported for MLA yet.
```
I am wondering why gated attention is currently not supported for MLA, and whether it might be possible in the future.

**Describe the solution you'd like**
Add support for gated attention in MLA

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.