Skip to main content

Flash Attention CUTE (CUDA Template Engine) implementation

Project description

FlashAttention-4 (CuTeDSL)

FlashAttention-4 is a CuTeDSL-based implementation of FlashAttention for Hopper and Blackwell GPUs.

Installation

pip install flash-attn-4

If you're on CUDA 13, install with the cu13 extra for best performance:

pip install "flash-attn-4[cu13]"

Usage

from flash_attn.cute import flash_attn_func, flash_attn_varlen_func

out = flash_attn_func(q, k, v, causal=True)

Development

git clone https://github.com/Dao-AILab/flash-attention.git
cd flash-attention
pip install -e "flash_attn/cute[dev]"       # CUDA 12.x
pip install -e "flash_attn/cute[dev,cu13]"  # CUDA 13.x (e.g. B200)
pytest tests/cute/

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

flash_attn_4-4.0.0b21.tar.gz (353.2 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

flash_attn_4-4.0.0b21-py3-none-any.whl (377.7 kB view details)

Uploaded Python 3

File details

Details for the file flash_attn_4-4.0.0b21.tar.gz.

File metadata

  • Download URL: flash_attn_4-4.0.0b21.tar.gz
  • Upload date:
  • Size: 353.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for flash_attn_4-4.0.0b21.tar.gz
Algorithm Hash digest
SHA256 1eaf7178aaf3c8d40c7162ee45c2d0165480bcb9495ddae66e80b3e7f1c17cda
MD5 b7ff7cd6940eac4e424146b31060fa6d
BLAKE2b-256 a21a36a1981aac29822763228805ed73669fb46afa865310c2273fc6f9849fe9

See more details on using hashes here.

Provenance

The following attestation bundles were made for flash_attn_4-4.0.0b21.tar.gz:

Publisher: publish-fa4.yml on Dao-AILab/flash-attention

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file flash_attn_4-4.0.0b21-py3-none-any.whl.

File metadata

  • Download URL: flash_attn_4-4.0.0b21-py3-none-any.whl
  • Upload date:
  • Size: 377.7 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for flash_attn_4-4.0.0b21-py3-none-any.whl
Algorithm Hash digest
SHA256 8ad0cdd22b279ef2ca8b4e43a4fb33b7dc78936d36c41788bad9d67372056949
MD5 3fbec80194b4480fe79da499fdce3bff
BLAKE2b-256 441c7ee1a5949e9b275eb62017d6fcc70e385db161b5341b686a29beb94ea181

See more details on using hashes here.

Provenance

The following attestation bundles were made for flash_attn_4-4.0.0b21-py3-none-any.whl:

Publisher: publish-fa4.yml on Dao-AILab/flash-attention

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page