xformers-0.0.33

xformers=0.0.33 compiled for vllm-0.10.1rc2
CC=12.0 (compute capability of NVIDIA GPUs - eg: RTX 5090, more: https://developer.nvidia.com/cuda-gpus)
CUDA=12.8.1
Commit hash=c159edc05ae5a0192ab0558e834b946155790371

vllm-0.10.1rc2-cc-12.0

vllm=vllm-0.10.1rc2 compiled for:
CC=12.0 (compute capability of NVIDIA GPUs - eg: RTX 5090, more: https://developer.nvidia.com/cuda-gpus)
CUDA=12.8.1
Commit hash=0583578f42fb23cc8a6d612e041c6be402551282

vllm-0.10.1.dev

vllm=vllm-0.10.1.dev compiled for:
CC=12.0 (compute capability of NVIDIA GPUs - eg: RTX 5090, more: https://developer.nvidia.com/cuda-gpus)
CUDA=12.8.1

torchvision-0.24.0-torch-2.9-cuda-12.8

torchvision=0.24.0 compiled for torch_2.9_cuda_12.8.1_cc_120
CC=12.0 (compute capability of NVIDIA GPUs - eg: RTX 5090, more: https://developer.nvidia.com/cuda-gpus)
CUDA=12.8.1
Commit hash: 48b2f9fb0b08713b853410624391a4a560e1ef3d

torch_2.9_cuda_12.8.1_cc_120

torch=2.9.0 compiled for:
CC=12.0 (compute capability of NVIDIA GPUs - eg: RTX 5090, more: https://developer.nvidia.com/cuda-gpus)
CUDA=12.8.1
Commit hash=4fd761fecca86693c8cd7755ea6d13d61bb94090

flashinfer-0.2.14.post1-cu128-torch29

CUDA=12.8.1
CC=12.0 (compute capability of NVIDIA GPUs)
Commit hash: 287bb64a376cbb110818a4a5aaeb5a7aa4a6ebb8

flashinfer-cu126-torch270

Compiled wheel file for Blackwell and previous architecture NVIDIA GPUs (8.9, 9.0, 10.0, 12.0 according to the table via link:
https://developer.nvidia.com/cuda-gpus

python 3.11
torch 2.7.0
flashinfer from the github
cuda 12.6

flash-attn-cp311-torch-270-cu128

Compiled wheel file for Blackwell and previous architecture NVIDIA GPUs (8.9, 12.0 according to the table via link:
https://developer.nvidia.com/cuda-gpus

python 3.11
torch 2.7.0
flash-attn 2.7.4.post1
cuda 12.8

flash-attn-cp311-torch-270-cu126

Compiled wheel file for Blackwell and previous architecture NVIDIA GPUs (8.9, 9.0, 12.0 according to the table via link:
https://developer.nvidia.com/cuda-gpus

python 3.11
torch 2.7.0
flash-attn 2.7.4.post1
cuda 12.6