bmilde

Python2/python3 got probably mixed up in the install script, it's probably best to point the install scripts to your python3 install manually, see also https://github.com/uhh-lt/kaldi-model-server

Thanks for the suggestion! Would you have a good idea on how to solve versioning for the kaldi dependency? Basically each pykaldi version needs to be compiled to a specific...

Please give more details of your system (OS/distro, g++ version, numpy version, etc) I'd recommend using a virtual environment for Python to install everything. You can also try with Python...

Take a look in exp/model_l4_c320/log/ and search for the training log of iteration 25. There might be some clues in the log why the training failed. You can also try...

> It is not a great solution, but we’ve used it in the past to find utterances where the transcriptions were completely off (for example, something that had more symbols...

Thanks! Ultimately I've added this to train-ctc-parallel.cc, after the check for too long sequences ("has too many frames; ignoring:") : // Check that we have enough frames to align them...

Reusing the CuRand object does help! I'm guessing that seeding the CuRand object was happening on the host, not on the GPU. However my quick fix will only work, if...

Just a FYI, you don't need: ./check_dependencies.sh ./install_protobuf.sh ./install_clif.sh If you are installing from a whl package

done

I've encountered this as well with the whisper commandline vs. using whisper from a python script (both have different defaults), see here: https://github.com/openai/whisper/discussions/591 The default parameters that the python whisper...