colabfold-legacy-kernels
Prebuilt CUDA kernels to make ColabFold fast on Volta and Turning NVIDIA GPUs. This package has three kernels that replace the Ampere+ Pallas/Triton kernels: Attention, layer norm and the gated dual projection.
import colabfold_legacy_kernels as clk
clk.available(cc=70) # True if the wheel has kernels for sm_70
clk.library_path("attention", cc=70) # path of the shared library
clk.symbol("attention", cc=70) # XLA FFI target name to register
To build the kernels, run the build script. It needs nvcc and the XLA FFI headers, and fetches CUTLASS itself.
NVCC=/usr/local/cuda-12.9/bin/nvcc scripts/build_all.sh # sm_70 and sm_75
ARCH=75 scripts/build_kernels.sh # one architecture
1 maintainer
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distributions
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file colabfold_legacy_kernels-0.2.0-py3-none-manylinux_2_28_x86_64.whl.
File metadata
- Download URL: colabfold_legacy_kernels-0.2.0-py3-none-manylinux_2_28_x86_64.whl
- Upload date:
- Size: 5.0 MB
- Tags: Python 3, manylinux: glibc 2.28+ x86-64
- Uploaded using Trusted Publishing? Yes
- Uploaded via:
twine/7.0.0 CPython/3.13.14
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
d125c58d87c5a90f230d65b32e00fccdb35ba8d9ba13de7256cd7edb3a8634bc
|
|
| MD5 |
e68588eb654c47a3f1ed27adcaf7831c
|
|
| BLAKE2b-256 |
f8220a21c041a1858eb604f97effaf49982459fd8070842bf59281b87308077b
|
Provenance
The following attestation bundles were made for colabfold_legacy_kernels-0.2.0-py3-none-manylinux_2_28_x86_64.whl:
Publisher:
build-and-publish.yml on mirditalab/colabfold-legacy-kernels
-
Statement:
-
Statement type:
https://in-toto.io/Statement/v1 -
Predicate type:
https://docs.pypi.org/attestations/publish/v1 -
Subject name:
colabfold_legacy_kernels-0.2.0-py3-none-manylinux_2_28_x86_64.whl -
Subject digest:
d125c58d87c5a90f230d65b32e00fccdb35ba8d9ba13de7256cd7edb3a8634bc - Sigstore transparency entry: 2826575363
- Sigstore integration time:
-
Permalink:
mirditalab/colabfold-legacy-kernels@4b60d7b0afe3be82bee600a5db9db86c6811d603 -
Branch / Tag:
refs/tags/v0.2.0 - Owner: https://github.com/mirditalab
-
Access:
public
-
Token Issuer:
https://token.actions.githubusercontent.com -
Runner Environment:
github-hosted -
Publication workflow:
build-and-publish.yml@4b60d7b0afe3be82bee600a5db9db86c6811d603 -
Trigger Event:
push
-
Statement type:
