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 14, 05:30 UTC
fix type of sampling rate for encode_base64 (#24826)
Signed-off-by: co63oc <co63oc@users.noreply.github.com>
co63oc committed
Sep 15, 2025 at 00:17 UTC
1177dd53e95f2e65f253b1ac613ecf4768327e49
2 files changed
+2
-2
vllm/multimodal/audio.py
+1
-1
@@ -106,7 +106,7 @@ class AudioMediaIO(MediaIO[tuple[npt.NDArray, float]]):
106
def load_file(self, filepath: Path) -> tuple[npt.NDArray, float]:
107
return librosa.load(filepath, sr=None)
108
109
- def encode_base64(self, media: tuple[npt.NDArray, float]) -> str:
109
+ def encode_base64(self, media: tuple[npt.NDArray, int]) -> str:
110
audio, sr = media
111
112
with BytesIO() as buffer:
vllm/multimodal/utils.py
+1
-1
@@ -310,7 +310,7 @@ class MediaConnector:
310
311
def encode_audio_base64(
312
audio: np.ndarray,
313
- sampling_rate: float,
313
+ sampling_rate: int,
314
) -> str:
315
"""Encode audio as base64."""
316
audio_io = AudioMediaIO()