Whisper 模型下载指南
Whisper 是 OpenAI 开源的语音识别模型，支持多语言转录。以下是使用 Hugging Face Hub 下载从 base 到 large-v3 所有版本的完整实现方案：
安装依赖
pip install transformers huggingface_hub
Python 下载代码
from huggingface_hub import snapshot_download
model_versions = [
"openai/whisper-base",
"openai/whisper-small",
"openai/whisper-medium",
"openai/whisper-large",
"openai/whisper-large-v2",
"openai/whisper-large-v3"
]
download_path = "./whisper_models"
for model_id in model_versions:
version_dir = f"{download_path}/{model_id.split('/')[-1]}"
snapshot_download(
repo_id=model_id,
local_dir=version_dir,
local_dir_use_symlinks=False,
resume_download=True
)
print(f"✅ {model_id} 下载完成")
关键参数说明
local_dir：自定义存储路径resume_download=True：支持断点续传
文件结构示例：
whisper_models/
├── whisper-base
├── whisper-small
├── whisper-medium
├── whisper-large
├── whisper-large-v2
└── whisper-large-v3
