00INDEX
Follow
Highlights
- Pro
Pinned
473 contributions in the last year
Activity overview
Contributed to
OpenLMLab/collie,
open-nlplab/fastIE,
00INDEX/Uni-app-WebProject
and 11 other
repositories
Contribution activity
April 2023
Created 81 commits in 2 repositories
Created 1 repository
- 00INDEX/collie Python
Opened 19 pull requests in 1 repository
OpenLMLab/collie
18
merged
1
closed
- Update llama_colossalai.py
- update requirements.txt
- fix save_state_dict
- modified save_state_dict
- fix save_state_dict
- fix save_state_dict
- add speed benchmark for zero
- adjust speed benchmark for colossalai
- add speed benchmark for colossalai api
- Add convert_model: convert models in Collie format to HF or RAW format
- Add Colossal-AI tensor parallel support
- Add more example for Colossal-AI
- Add more Colossal-AI examples
- Add runnable examples
- add examples
- fix README.md
- add README.md & setup.py
- add README.md & setup.py
- 实现colossal-Ai pipeline版本的llama
Reviewed 1 pull request in 1 repository
OpenLMLab/collie
1 pull request
Created an issue in HazyResearch/flash-attention that received 5 comments
[Question] What is the difference between FlashAttention and Memory Efficient Attention in xformers?
The paper I recently read, SELF-ATTENTION DOES NOT NEED O(n2) MEMORY, uses kernel fusion to overcome large GPU memory usage (to be more specificall…
5
comments