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, 03:30 UTC
[CI Bugfix] Fix pre-commit failures on main (#20502)
Michael Goin committed
Jul 5, 2025 at 06:17 UTC
c108781c856eb5b89e927b1bf667a724e405b2ac
2 files changed
+4
-2
tests/kernels/moe/test_cutlass_grouped_gemm.py
+1
@@ -1,4 +1,5 @@
1
# SPDX-License-Identifier: Apache-2.0
2
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
3
4
# DeepGEMM Style Cutlass Grouped GEMM Test
5
# See https://github.com/deepseek-ai/DeepGEMM/blob/main/tests/test_core.py
vllm/multimodal/utils.py
+3
-2
@@ -48,9 +48,10 @@ class MediaConnector:
48
media_io_kwargs: Additional args passed to process media
49
inputs, keyed by modalities. For example,
50
to set num_frames for video, set
51
- `--media-io-kwargs '{"video": {"num_frames": 40} }'`
51
+ `--media-io-kwargs '{"video":{"num_frames":40}}'`
52
connection: HTTP connection client to download media contents.
53
- allowed_local_media_path: A local directory to load media files from.
53
+ allowed_local_media_path: A local directory to load media files
54
+ from.
55
"""
56
super().__init__()
57