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 15, 12:15 UTC
Update test requirements to CUDA 12.8 (#17576)
Signed-off-by: 22quinn <33176974+22quinn@users.noreply.github.com>
22quinn committed
May 2, 2025 at 21:40 UTC
d47b605eca42d66731283d7aecfea2b62d047402
2 files changed
+17
-17
.pre-commit-config.yaml
+1
-1
@@ -46,7 +46,7 @@ repos:
46
rev: 0.6.17
47
hooks:
48
- id: pip-compile
49
- args: [requirements/test.in, -o, requirements/test.txt, --index-strategy, unsafe-best-match]
49
+ args: [requirements/test.in, -o, requirements/test.txt, --index-strategy, unsafe-best-match, --torch-backend, cu128]
50
files: ^requirements/test\.(in|txt)$
51
- repo: local
52
hooks:
requirements/test.txt
+16
-16
@@ -1,5 +1,5 @@
1
# This file was autogenerated by uv via the following command:
2
-# uv pip compile requirements/test.in -o requirements/test.txt --index-strategy unsafe-best-match
2
+# uv pip compile requirements/test.in -o requirements/test.txt --index-strategy unsafe-best-match --torch-backend cu128
3
absl-py==2.1.0
4
# via rouge-score
5
accelerate==1.0.1
@@ -349,28 +349,28 @@ numpy==1.26.4
349
# transformers
350
# tritonclient
351
# vocos
352
-nvidia-cublas-cu12==12.6.4.1
352
+nvidia-cublas-cu12==12.8.3.14
353
# via
354
# nvidia-cudnn-cu12
355
# nvidia-cusolver-cu12
356
# torch
357
-nvidia-cuda-cupti-cu12==12.6.80
357
+nvidia-cuda-cupti-cu12==12.8.57
358
# via torch
359
-nvidia-cuda-nvrtc-cu12==12.6.77
359
+nvidia-cuda-nvrtc-cu12==12.8.61
360
# via torch
361
-nvidia-cuda-runtime-cu12==12.6.77
361
+nvidia-cuda-runtime-cu12==12.8.57
362
# via torch
363
-nvidia-cudnn-cu12==9.5.1.17
363
+nvidia-cudnn-cu12==9.7.1.26
364
# via torch
365
-nvidia-cufft-cu12==11.3.0.4
365
+nvidia-cufft-cu12==11.3.3.41
366
# via torch
367
-nvidia-cufile-cu12==1.11.1.6
367
+nvidia-cufile-cu12==1.13.0.11
368
# via torch
369
-nvidia-curand-cu12==10.3.7.77
369
+nvidia-curand-cu12==10.3.9.55
370
# via torch
371
-nvidia-cusolver-cu12==11.7.1.2
371
+nvidia-cusolver-cu12==11.7.2.55
372
# via torch
373
-nvidia-cusparse-cu12==12.5.4.2
373
+nvidia-cusparse-cu12==12.5.7.53
374
# via
375
# nvidia-cusolver-cu12
376
# torch
@@ -378,13 +378,13 @@ nvidia-cusparselt-cu12==0.6.3
378
# via torch
379
nvidia-nccl-cu12==2.26.2
380
# via torch
381
-nvidia-nvjitlink-cu12==12.6.85
381
+nvidia-nvjitlink-cu12==12.8.61
382
# via
383
# nvidia-cufft-cu12
384
# nvidia-cusolver-cu12
385
# nvidia-cusparse-cu12
386
# torch
387
-nvidia-nvtx-cu12==12.6.77
387
+nvidia-nvtx-cu12==12.8.55
388
# via torch
389
opencv-python-headless==4.11.0.86
390
# via
@@ -687,7 +687,7 @@ tomli==2.2.1
687
# via schemathesis
688
tomli-w==1.2.0
689
# via schemathesis
690
-torch==2.7.0
690
+torch==2.7.0+cu128
691
# via
692
# -r requirements/test.in
693
# accelerate
@@ -705,12 +705,12 @@ torch==2.7.0
705
# torchvision
706
# vector-quantize-pytorch
707
# vocos
708
-torchaudio==2.7.0
708
+torchaudio==2.7.0+cu128
709
# via
710
# -r requirements/test.in
711
# encodec
712
# vocos
713
-torchvision==0.22.0
713
+torchvision==0.22.0+cu128
714
# via
715
# -r requirements/test.in
716
# timm