lots of issues with internal server error

[Eve-146T]

version: 0.7.0

config:

docker run \
  --restart=always \
  --detach \
  --publish 8026:8000 \
  --name speaches \
  --env WHISPER__MODEL="Systran/faster-whisper-large-v2" \
  --env WHISPER__USE_BATCHED_MODE="true" \
  --volume ~/.cache/huggingface:/root/.cache/huggingface \
  --gpus=all \
  ghcr.io/speaches-ai/speaches:latest-cuda

small audio files work fine, like a few seconds, but if you upload a larger file it errors out:

Error
Server error '500 Internal Server Error' for url 'http://localhost:8026/v1/audio/transcriptions' For more information check: https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/500

it not only fails in the web interface, but also in the other apps that i built that rely on the server.

This was not the case in the "old" faster-whisper-server"

[fedirz]

I'm going to need either the file you were trying to transcribe or additional logs to debug this issue.

[GTez]

On a fresh docker compose install I get these errors

File "/home/ubuntu/speaches/speaches/hf_utils.py", line 226, in get_kokoro_model_path
2025-02-19 22:09:46,888:DEBUG:httpcore.http11:atrace:85:receive_response_headers.complete return_value=(b'HTTP/1.1', 500, b'Internal Server Error', [(b'date', b'Wed, 19 Feb 2025 22:09:46 GMT'), (b'server', b'uvicorn'), (b'content-length', b'21'), (b'content-type', b'text/plain; charset=utf-8')])
2025-02-19 22:09:46,888:INFO:httpx:_send_single_request:1786:HTTP Request: POST http://10.0.8.50:8000/v1/audio/speech "HTTP/1.1 500 Internal Server Error"
    raise ValueError(f"Could not find {file_name} file for 'hexgrad/Kokoro-82M' model")
ValueError: Could not find kokoro-v0_19.onnx file for 'hexgrad/Kokoro-82M' model
2025-02-19 22:09:46,888:DEBUG:httpcore.http11:atrace:85:receive_response_body.started request=<Request [b'POST']>
2025-02-19 22:09:46,888:DEBUG:httpcore.http11:atrace:85:receive_response_body.complete
2025-02-19 22:09:46,888:DEBUG:httpcore.http11:atrace:85:response_closed.started
2025-02-19 22:09:46,888:DEBUG:httpcore.http11:atrace:85:response_closed.complete
2025-02-19 22:09:46,888:DEBUG:httpcore.connection:atrace:85:close.started
2025-02-19 22:09:46,888:DEBUG:httpcore.connection:atrace:85:close.complete
2025-02-19 22:09:46,888:DEBUG:openai._base_client:_request:1618:HTTP Request: POST http://10.0.8.50:8000/v1/audio/speech "500 Internal Server Error"
2025-02-19 22:09:46,888:DEBUG:openai._base_client:_request:1625:Encountered httpx.HTTPStatusError
Traceback (most recent call last):
  File "/home/ubuntu/speaches/.venv/lib/python3.12/site-packages/openai/_base_client.py", line 1623, in _request
    response.raise_for_status()
  File "/home/ubuntu/speaches/.venv/lib/python3.12/site-packages/httpx/_models.py", line 763, in raise_for_status
    raise HTTPStatusError(message, request=request, response=self)
httpx.HTTPStatusError: Server error '500 Internal Server Error' for url 'http://10.0.8.50:8000/v1/audio/speech'
For more information check: https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/500

Edit: Looks like I needed to download the models: per this https://speaches.ai/usage/text-to-speech/

Edit 2 - The following had an error:

docker exec -it speaches curl --location --output /home/ubuntu/.cache/huggingface/hub/models--hexgrad--Kokoro-82M/snapshots/$KOKORO_REVISION/voices.bin https://github.com/thewh1teagle/kokoro-onnx/releases/download/model-files/voices.bin

curl: (77) error setting certificate file: /etc/ssl/certs/ca-certificates.crt

And so I needed to pass the -k flag to curl - And audio output!

[LJ-Software]

Thank you for mentioning the -k flag @GTez that helped me get unblocked on this.
I ran into the same certificate error when trying to download the voice models.