423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-12-27 09:10:35 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix

Browse Source
This commit is contained in:
hiyouga
2025-12-27 08:05:25 +08:00
parent 5f3e5eb5f8
commit 42b436a4fc
2 changed files with 13 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
pyproject.toml
View File

@@ -40,6 +40,7 @@ dependencies = [
# core deps # core deps
"torch>=2.4.0", "torch>=2.4.0",
"torchvision>=0.19.0", "torchvision>=0.19.0",
"torchaudio>=2.4.0",
"transformers>=4.49.0,<=4.56.2,!=4.52.0; python_version < '3.10'", "transformers>=4.49.0,<=4.56.2,!=4.52.0; python_version < '3.10'",
"transformers>=4.49.0,<=4.57.1,!=4.52.0,!=4.57.0; python_version >= '3.10'", "transformers>=4.49.0,<=4.57.1,!=4.52.0,!=4.57.0; python_version >= '3.10'",
"datasets>=2.16.0,<=4.0.0", "datasets>=2.16.0,<=4.0.0",
@@ -63,6 +64,7 @@ dependencies = [
"hf-transfer", "hf-transfer",
"safetensors", "safetensors",
# python # python
"av",
"fire", "fire",
"omegaconf", "omegaconf",
"packaging", "packaging",
@@ -72,10 +74,7 @@ dependencies = [
# api # api
"uvicorn", "uvicorn",
"fastapi", "fastapi",
"sse-starlette", "sse-starlette"
# media
"av",
"librosa"
] ]
[project.optional-dependencies] [project.optional-dependencies]

21
src/llamafactory/data/mm_plugin.py
View File

@@ -26,6 +26,7 @@ from typing import TYPE_CHECKING, BinaryIO, Literal, Optional, TypedDict, Union
import numpy as np import numpy as np
import torch import torch
import torchaudio
from transformers.image_utils import get_image_size, is_valid_image, to_numpy_array from transformers.image_utils import get_image_size, is_valid_image, to_numpy_array
from transformers.models.mllama.processing_mllama import ( from transformers.models.mllama.processing_mllama import (
convert_sparse_cross_attention_mask_to_dense, convert_sparse_cross_attention_mask_to_dense,
@@ -34,16 +35,7 @@ from transformers.models.mllama.processing_mllama import (
from typing_extensions import NotRequired, override from typing_extensions import NotRequired, override
from ..extras.constants import AUDIO_PLACEHOLDER, IGNORE_INDEX, IMAGE_PLACEHOLDER, VIDEO_PLACEHOLDER from ..extras.constants import AUDIO_PLACEHOLDER, IGNORE_INDEX, IMAGE_PLACEHOLDER, VIDEO_PLACEHOLDER
from ..extras.packages import ( from ..extras.packages import is_pillow_available, is_pyav_available, is_transformers_version_greater_than
is_librosa_available,
is_pillow_available,
is_pyav_available,
is_transformers_version_greater_than,
)
if is_librosa_available():
import librosa
if is_pillow_available(): if is_pillow_available():
@@ -316,7 +308,14 @@ class MMPluginMixin:
results, sampling_rates = [], [] results, sampling_rates = [], []
for audio in audios: for audio in audios:
if not isinstance(audio, np.ndarray): if not isinstance(audio, np.ndarray):
audio, sampling_rate = librosa.load(audio, sr=sampling_rate) audio, sr = torchaudio.load(audio)
if audio.shape[0] > 1:
audio = audio.mean(dim=0, keepdim=True)
if sr != sampling_rate:
audio = torchaudio.functional.resample(audio, sr, sampling_rate)
audio = audio.squeeze(0).numpy()
results.append(audio) results.append(audio)
sampling_rates.append(sampling_rate) sampling_rates.append(sampling_rate)