#7168 Error during speech recognition stage[Huggingface_ASR(Built-in)] Automatic model download failed. Click [Help Docs] for

103.169**1 Posted at: 1 hour ago

Error during speech recognition stage[Huggingface_ASR(Built-in)] Automatic model download failed. Click [Help Docs] for manual download instructions:

Traceback (most recent call last):

File "httpx\_transports\default.py", line 101, in map_httpcore_exceptions

File "httpx\_transports\default.py", line 127, in iter

File "httpcore\_sync\connection_pool.py", line 367, in iter

File "httpcore\_sync\connection_pool.py", line 363, in iter

File "httpcore\_sync\http11.py", line 349, in iter

File "httpcore\_sync\http11.py", line 341, in iter

File "httpcore\_sync\http11.py", line 210, in _receive_response_body

File "httpcore\_sync\http11.py", line 220, in _receive_event

File "contextlib.py", line 153, in exit

File "httpcore\_exceptions.py", line 14, in map_exceptions

httpcore.RemoteProtocolError: peer closed connection without sending complete message body (received 709917743 bytes, expected 2508311120)

The above e
......
le "httpx\_models.py", line 897, in iter_bytes

File "httpx\_models.py", line 951, in iter_raw

File "httpx\_client.py", line 153, in iter

File "httpx\_transports\default.py", line 126, in iter

File "contextlib.py", line 153, in exit

File "httpx\_transports\default.py", line 118, in map_httpcore_exceptions

httpx.RemoteProtocolError: peer closed connection without sending complete message body (received 709917743 bytes, expected 2508311120)

During handling of the above exception, another exception occurred:

Traceback (most recent call last):

File "videotrans\task\job.py", line 35, in run

File "videotrans\task\job.py", line 102, in process_task

File "videotrans\task\_stage_recogn.py", line 65, in recogn

File "videotrans\recognition\__init__.py", line 228, in run

File "videotrans\recognition\_base.py", line 95, in run

File "videotrans\recognition\_huggingface.py", line 28, in _download

File "videotrans\util\help_down.py", line 168, in check_and_down_hf

videotrans.configure.excepts.DownloadModelsError: Automatic model download failed. Click [Help Docs] for manual download instructions:

[TransCreate]翻译视频: self.is_copy_video=False,self.video_codec_num=264,self.is_audio_trans=True,self.video_info={'video_fps': 30, 'r_frame_rate': 30, 'video_codec_name': '', 'audio_codec_name': 'mp3', 'width': 0, 'height': 0, 'time': 1510295, 'streams_len': 1, 'streams_audio': 1, 'video_streams': 0, 'color': 'yuv420p'}
[TaskCfgVTT]当前工作模式: 转录并翻译字幕
原始输入文件名: E:/15331461171960.mp3,
输出结果保存到文件夹: E:/video/work/15331461171960-mp3,
临时文件夹: E:/video/pyVideoTrans/tmp/19736/5d8b33779e
已选 清理已存在
未启用CUDA加速
未选 降噪
STT Service:Huggingface_ASR(Built-in), nvidia/parakeet-tdt-0.6b-v3, 发音语言: Russian, 断句方式:默认断句
TranslateSrt:Google Translate (Free),原始语言:Russian,目标语言:Bengali, 已选 SendSRT
SubFormat: 单字幕


=
system:Windows-10-10.0.26200-SP0
version:v4.12
frozen:True
language:en_US
root_dir:E:/video/pyVideoTrans

1 Replies

#1 DeepSeek reply at 1 hour ago

The core cause is a broken/incomplete model download: the HuggingFace model file transfer was interrupted by the remote peer (peer closed connection without sending complete message body), so the nvidia/parakeet-tdt-0.6b-v3 model for the Huggingface_ASR(Built-in) channel was never fully downloaded and the ASR stage aborted with DownloadModelsError.

Fix it as follows:

  1. Delete the partially-downloaded model folder to avoid corrupted leftovers: E:/video/pyVideoTrans/models/models--nvidia--parakeet-tdt-0.6b-v3
  2. Manually download the full model from HuggingFace: https://huggingface.co/nvidia/parakeet-tdt-0.6b-v3/tree/main

    • Download all files in that repo (keep exact original filenames; some models have several large .safetensors/.bin shards).
  3. Place them into: E:/video/pyVideoTrans/models/models--nvidia--parakeet-tdt-0.6b-v3

    • Naming rule: model display name nvidia/parakeet-tdt-0.6b-v3 → folder models--nvidia--parakeet-tdt-0.6b-v3 (the / becomes --).
  4. Re-run the task. Do not check 清理已生成 if you want to keep already-processed files, but note the incomplete model folder must be removed first.
  5. If downloads keep failing, use a stable proxy/VPN, or switch to hf-mirror.com as the mirror. Alternatively pick a different ASR channel (e.g. faster-whisper(内置) with large-v3), which is also a good fit for Russian.

Note: your log shows 未启用CUDA加速 (CUDA acceleration not enabled) — for a model this size, CPU-only runs will be very slow; enabling CUDA (NVIDIA + CUDA 12.8 + cuDNN 9.11) is strongly recommended.

请查阅相关文档:

Thinking...
Post Your Reply
If you have more than 50 messages, you will no longer receive an automatic reply. Please ask a new question!
Open source and free maintenance is not easy. If this project is helpful to you, please consider making a small donation to help the project continue to maintain and update.

Dondate via ko-fi.com

Related resource