xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-03-16 10:17:13 +08:00
Code Issues Packages Projects Releases Wiki Activity

[BugFix] [FEAT] Enable fastsafetensors for ROCm platform (#28225)

Browse Source 
Signed-off-by: tjtanaa <tunjian.tan@embeddedllm.com>
This commit is contained in:
TJian 2025-11-20 23:34:11 +07:00 committed by GitHub
parent a2e9ebe9e2
commit 82b05b15e6
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
3 changed files with 5 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
requirements/rocm.txt
View File

@ -15,3 +15,4 @@ setuptools-scm>=8
runai-model-streamer[s3,gcs]==0.15.0
conch-triton-kernels==1.2.1
timm>=1.0.17
fastsafetensors @ git+https://github.com/foundation-model-stack/fastsafetensors.git@d6f998a03432b2452f8de2bb5cefb5af9795d459

3
tests/model_executor/model_loader/fastsafetensors_loader/test_fastsafetensors_loader.py
View File

@ -19,7 +19,8 @@ sampling_params = SamplingParams(temperature=0.8, top_p=0.95, seed=0)
@pytest.mark.skipif(
not current_platform.is_cuda(), reason="fastsafetensors requires CUDA/NVIDIA GPUs"
not current_platform.is_cuda_alike(),
reason="fastsafetensors requires NVIDIA/AMD GPUs",
)
def test_model_loader_download_files(vllm_runner):
with vllm_runner(test_model, load_format="fastsafetensors") as llm:

3
tests/model_executor/model_loader/fastsafetensors_loader/test_weight_utils.py
View File

@ -17,7 +17,8 @@ from vllm.platforms import current_platform
@pytest.mark.skipif(
not current_platform.is_cuda(), reason="fastsafetensors requires CUDA/NVIDIA GPUs"
not current_platform.is_cuda_alike(),
reason="fastsafetensors requires NVIDIA/AMD GPUs",
)
def test_fastsafetensors_model_loader():
with tempfile.TemporaryDirectory() as tmpdir: