not in the list of choices

[ccmsi]

Greetings and thanks for providing speaches.ai
I had a little initial drama getting things to work in Kubernetes (ie adding persistence for the /home/ubuntu/.cache/huggingface/hub and providing CA certs (proposal: replace /home/ubuntu/speaches/.venv/lib/python3.12/site-packages/certifi/cacert.pem with a symlink to /etc/ssl/certs/ca-certificates.crt ) but seems to be working now:

$ curl "$SPEACHES_BASE_URL/v1/models/Systran/faster-distil-whisper-small.en" -X POST
Model 'Systran/faster-distil-whisper-small.en' downloaded

$ curl "$SPEACHES_BASE_URL/v1/models/speaches-ai/Kokoro-82M-v1.0-ONNX" -X POST
Model 'speaches-ai/Kokoro-82M-v1.0-ONNX' downloaded

Things seem fine now in the UI. I can select models, voices, microphone picks up sound, can replay, can select sample rate.

However, when getting to action (submitting task) i get back the error message "xyz is not in the list of choices" with an empty list "[]" in both STT and TTS.

File "/home/ubuntu/speaches/.venv/lib/python3.12/site-packages/gradio/components/dropdown.py", line 206, in preprocess
raise Error(
gradio.exceptions.Error: 'Value: speaches-ai/Kokoro-82M-v1.0-ONNX is not in the list of choices: []'

things seem normal to me:
$ find /home/ubuntu/.cache/huggingface/hub/ -type f
/home/ubuntu/.cache/huggingface/hub/models--Systran--faster-distil-whisper-small.en/refs/main
/home/ubuntu/.cache/huggingface/hub/models--Systran--faster-distil-whisper-small.en/blobs/e6e037dbb16523ca03f49d94e8da549efd08519e
/home/ubuntu/.cache/huggingface/hub/models--Systran--faster-distil-whisper-small.en/blobs/91876762a536a746d268353c5cba57286e76b058
/home/ubuntu/.cache/huggingface/hub/models--Systran--faster-distil-whisper-small.en/blobs/1187de3982cdcf962a2fb8f797e429fb4651b875b18fe9ce50b58b52fc9072b7
/home/ubuntu/.cache/huggingface/hub/models--Systran--faster-distil-whisper-small.en/blobs/b9bc6c16dcfbfc54ab235b55e079d4311f97edf5
/home/ubuntu/.cache/huggingface/hub/models--Systran--faster-distil-whisper-small.en/blobs/d7cd7ef030d38ff11df010adc96a29e3a0222151
/home/ubuntu/.cache/huggingface/hub/models--Systran--faster-distil-whisper-small.en/blobs/cab338fd7cd3a2f484ee8ece80d3c63964d47c4a
/home/ubuntu/.cache/huggingface/hub/.locks/models--Systran--faster-distil-whisper-small.en/d7cd7ef030d38ff11df010adc96a29e3a0222151.lock
/home/ubuntu/.cache/huggingface/hub/.locks/models--Systran--faster-distil-whisper-small.en/1187de3982cdcf962a2fb8f797e429fb4651b875b18fe9ce50b58b52fc9072b7.lock
/home/ubuntu/.cache/huggingface/hub/.locks/models--Systran--faster-distil-whisper-small.en/e6e037dbb16523ca03f49d94e8da549efd08519e.lock
/home/ubuntu/.cache/huggingface/hub/.locks/models--Systran--faster-distil-whisper-small.en/b9bc6c16dcfbfc54ab235b55e079d4311f97edf5.lock
/home/ubuntu/.cache/huggingface/hub/.locks/models--Systran--faster-distil-whisper-small.en/cab338fd7cd3a2f484ee8ece80d3c63964d47c4a.lock
/home/ubuntu/.cache/huggingface/hub/.locks/models--Systran--faster-distil-whisper-small.en/91876762a536a746d268353c5cba57286e76b058.lock
/home/ubuntu/.cache/huggingface/hub/.locks/models--speaches-ai--Kokoro-82M-v1.0-ONNX/bca610b8308e8d99f32e6fe4197e7ec01679264efed0cac9140fe9c29f1fbf7d.lock
/home/ubuntu/.cache/huggingface/hub/.locks/models--speaches-ai--Kokoro-82M-v1.0-ONNX/7d5df8ecf7d4b1878015a32686053fd0eebe2bc377234608764cc0ef3636a6c5.lock
/home/ubuntu/.cache/huggingface/hub/.locks/models--speaches-ai--Kokoro-82M-v1.0-ONNX/bd5c4d72cead61a691b2253858ce337e3848770f.lock
/home/ubuntu/.cache/huggingface/hub/models--speaches-ai--Kokoro-82M-v1.0-ONNX/refs/main
/home/ubuntu/.cache/huggingface/hub/models--speaches-ai--Kokoro-82M-v1.0-ONNX/blobs/bca610b8308e8d99f32e6fe4197e7ec01679264efed0cac9140fe9c29f1fbf7d
/home/ubuntu/.cache/huggingface/hub/models--speaches-ai--Kokoro-82M-v1.0-ONNX/blobs/7d5df8ecf7d4b1878015a32686053fd0eebe2bc377234608764cc0ef3636a6c5
/home/ubuntu/.cache/huggingface/hub/models--speaches-ai--Kokoro-82M-v1.0-ONNX/blobs/bd5c4d72cead61a691b2253858ce337e3848770f

Thanks for helping out.

[ccmsi]

I have to correct myself slightly, unfortunately not for the better in terms of finding the issue.
While usually not, in the odd case, it just happened that STT... worked once, without changing anything. Still the same image/pod running. And then again it did not. Even not showing options in the drop downs when switching tasks (STT/TTS).
After which a hard reload (in browser) brought back the options in dropdowns after which i tried TTS and it worked.. once.
There may be two issues at play.
Since this is probably not reproducible on your end i'm mainly asking for advice on where and what for i should be investigating.