Update README with CUDA 13.0 build info

Fix flash_attn_config import to use relative import in cu130 x86_64 builds

Update flash_attn_interface.py and flash_attn_config.py for cu130 x86_64 builds

Update cu130 x86_64 binaries with correct CUDA 13.0 build

Add torch 2.11 and 2.12 builds (ABI-compatible with 2.9+)

Upload 56 files

Delete build

Upload 8 files

Update README.md

Upload 8 files

Update README.md

Upload 16 files

Create config.json

update readme

Upload 8 files

Upload __init__.py

Upload __init__.py

Upload 6 files

Upload README.md

Update build/torch28-cxx11-cu126-x86_64-linux/flash_attention_3/flash_attn_interface.py

Update build/torch28-cxx11-cu128-x86_64-linux/flash_attention_3/flash_attn_interface.py

Upload __init__.py

Upload __init__.py

Upload 8 files

Delete build

Upload 4 files

Upload 5 files

Delete build

Upload README

Upload 4 files

initial commit