argument error when doing vad_filter
asafarevich opened this issue · 1 comments
asafarevich commented
whisper-ctranslate2 --live_transcribe True --language en --device cpu --model small --vad_filter True
It will listen, but when it comes to transcribing it will crash with this error
Traceback (most recent call last):
File "/home/hackerman/anaconda3/envs/whisper/bin/whisper-ctranslate2", line 8, in <module>
sys.exit(main())
File "/home/hackerman/anaconda3/envs/whisper/lib/python3.9/site-packages/src/whisper_ctranslate2/whisper_ctranslate2.py", line 498, in main
Live(
File "/home/hackerman/anaconda3/envs/whisper/lib/python3.9/site-packages/src/whisper_ctranslate2/live.py", line 163, in inference
self.listen()
File "/home/hackerman/anaconda3/envs/whisper/lib/python3.9/site-packages/src/whisper_ctranslate2/live.py", line 159, in listen
self.process()
File "/home/hackerman/anaconda3/envs/whisper/lib/python3.9/site-packages/src/whisper_ctranslate2/live.py", line 134, in process
result = self.transcribe.inference(
File "/home/hackerman/anaconda3/envs/whisper/lib/python3.9/site-packages/src/whisper_ctranslate2/transcribe.py", line 128, in inference
segments, info = self.model.transcribe(
File "/home/hackerman/anaconda3/envs/whisper/lib/python3.9/site-packages/faster_whisper/transcribe.py", line 252, in transcribe
speech_chunks = get_speech_timestamps(audio, vad_parameters)
File "/home/hackerman/anaconda3/envs/whisper/lib/python3.9/site-packages/faster_whisper/vad.py", line 94, in get_speech_timestamps
speech_prob, state = model(chunk, state, sampling_rate)
File "/home/hackerman/anaconda3/envs/whisper/lib/python3.9/site-packages/faster_whisper/vad.py", line 288, in __call__
out, h, c = self.session.run(None, ort_inputs)
File "/home/hackerman/anaconda3/envs/whisper/lib/python3.9/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py", line 217, in run
return self._sess.run(output_names, input_feed, run_options)
onnxruntime.capi.onnxruntime_pybind11_state.InvalidArgument: [ONNXRuntimeError] : 2 : INVALID_ARGUMENT : Unexpected input data type. Actual: (tensor(double)) , expected: (tensor(float))
jordimas commented
This should be fixed in version 0.2.7 just released. Thanks for reporting