squall/flash-attention
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
1 Commit 1 Branch 0 Tags 7.9 MiB
Python 48.6%
C++ 41.1%
Cuda 10.1%
Dockerfile 0.1%
1fcbe6f0d0
Go to file
Tri Dao 1fcbe6f0d0 First release
2022-05-20 14:21:58 -07:00
csrc/stream_attn First release 2022-05-20 14:21:58 -07:00
bert_padding.py First release 2022-05-20 14:21:58 -07:00
README.md First release 2022-05-20 14:21:58 -07:00
rotary.py First release 2022-05-20 14:21:58 -07:00
stream_attn_interface.py First release 2022-05-20 14:21:58 -07:00
stream_blocksparse_attn_interface.py First release 2022-05-20 14:21:58 -07:00
streaming_attention.py First release 2022-05-20 14:21:58 -07:00
streaming_blocksparse_attention.py First release 2022-05-20 14:21:58 -07:00

README.md

Alpha release of FlashAttention.

To compile:

cd csrc/stream_attn
python setup.py install

Interface: streaming_attention.py

Contact: trid@stanford.edu