[Misc] Optional installation of audio related packages (#8063)

This commit is contained in:
Roger Wang
2024-09-01 14:46:57 -07:00
committed by GitHub
parent 5231f0898e
commit 5b86b19954
6 changed files with 29 additions and 10 deletions
+5 -1
View File
@@ -8,7 +8,6 @@ from functools import lru_cache
from typing import (Iterable, List, Literal, Mapping, Optional, Tuple,
TypedDict, Union, cast)
import librosa
import numpy as np
import torch
import torch.utils.checkpoint
@@ -107,6 +106,11 @@ def input_mapper_for_ultravox(ctx: InputContext, data: object):
feature_extractor = whisper_feature_extractor(ctx)
if sr != feature_extractor.sampling_rate:
try:
import librosa
except ImportError:
raise ImportError(
"Please install vllm[audio] for audio support.") from None
audio = librosa.resample(audio,
orig_sr=sr,
target_sr=feature_extractor.sampling_rate)
+17 -3
View File
@@ -1,11 +1,9 @@
import base64
from functools import lru_cache
from io import BytesIO
from typing import List, Optional, Tuple, TypeVar, Union
from typing import Any, List, Optional, Tuple, TypeVar, Union
import librosa
import numpy as np
import soundfile
from PIL import Image
from vllm.connections import global_http_connection
@@ -73,10 +71,22 @@ async def async_fetch_image(image_url: str,
return image.convert(image_mode)
def try_import_audio_packages() -> Tuple[Any, Any]:
try:
import librosa
import soundfile
except ImportError:
raise ImportError(
"Please install vllm[audio] for audio support.") from None
return librosa, soundfile
def fetch_audio(audio_url: str) -> Tuple[np.ndarray, Union[int, float]]:
"""
Load audio from a URL.
"""
librosa, _ = try_import_audio_packages()
if audio_url.startswith("http"):
audio_bytes = global_http_connection.get_bytes(
audio_url, timeout=VLLM_AUDIO_FETCH_TIMEOUT)
@@ -95,6 +105,8 @@ async def async_fetch_audio(
"""
Asynchronously fetch audio from a URL.
"""
librosa, _ = try_import_audio_packages()
if audio_url.startswith("http"):
audio_bytes = await global_http_connection.async_get_bytes(
audio_url, timeout=VLLM_AUDIO_FETCH_TIMEOUT)
@@ -123,6 +135,8 @@ def encode_audio_base64(
sampling_rate: int,
) -> str:
"""Encode audio as base64."""
_, soundfile = try_import_audio_packages()
buffered = BytesIO()
soundfile.write(buffered, audio, sampling_rate, format="WAV")