From where i must download new version of pretrained model?
My deepspeech version is v0.6.0-alpha.15-0-gb3787ee
Alpha versions don’t have pre-trained models. You need to use v0.5.1 or wait for v0.6.0 stable release.
1 Like
Oh, i understand. Thank you!
Sorry, Can I ask one more question?
I downgraded my deepspeech to version 0.5.1, downloaded pretrained model for 0.5.1. But i have new problem!
I want to use their implementation of "Microphone VAD Streaming"
(https://github.com/mozilla/DeepSpeech/tree/master/examples/mic_vad_streaming). I installed all requirements and when I ran python mic_vad_streaming.py -m models/output_graph.pb -l models/lm.binary -t models/trie it dropped me error TypeError: Required argument 'aNContext' (pos 3) not found. I looked at the files and noted that there isn’t few arguments for deepspeech.Model() (not only aNContext). Maybe you know which arguments must be there by default and what values must they have.
If you’re using version v0.5.1, then you need to use the example code for v0.5.1: https://github.com/mozilla/DeepSpeech/tree/v0.5.1/examples/mic_vad_streaming
1 Like
Oh, thank you very much! I just forgot about this example!