I was fine tuning release 0.3.0 model from the checkpoint that Deepspeech team has provided using Common voice dataset but I forgot to mention a directory for ‘export_dir’ option of Deepspeech.py. Now, the training is complete but I don’t have a model but there are some new meta, data and index checkpoint files in the checkpoint directory. Is there a way I could still get the model?
Just re-run with the same settings for model / checkpoint, with --notrain and do not forget --export_dir, and that should be okay