A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 1, 18:45 UTC
[Dependency] Remove stale cuDNN frontend upper bound (#42599)
Signed-off-by: Mohammad Miadh Angkad <176301910+mmangkad@users.noreply.github.com>
Mohammad Miadh Angkad committed
Jun 7, 2026 at 16:09 UTC
66ecfd05681b1b5bf8c10ace9c8ed96cc5ac7659
1 file changed
+1
-3
requirements/cuda.txt
+1
-3
@@ -13,9 +13,7 @@ flashinfer-python==0.6.12
13
flashinfer-cubin==0.6.12
14
apache-tvm-ffi==0.1.9
15
tilelang==0.1.9
16
-# Cap nvidia-cudnn-frontend (transitive dep of flashinfer) due to
17
-# breaking changes in 1.19.0
18
-nvidia-cudnn-frontend>=1.13.0,<1.19.0
16
+nvidia-cudnn-frontend>=1.19.1
17
18
# Required for faster safetensors model loading
19
fastsafetensors >= 0.3.2