Add Silero Speech-To-Text models #153
Conversation
|
Deploy preview for pytorch-hub-preview ready! Built with commit 78f39ca |
|
Maybe I did something wrong, but I cannot see our model's page here |
|
Hi, @ailzhang @bertmaher @wconstab could you please help with assigning a correct person to review the submission? |
There was a problem hiding this comment.
@snakers4
Thanks for contributing!
If you move your md file out of docs you should be able to see it on the webpage.
Also once you do that, code snippet in the the markdown will be executed in the CI so you probably also need a working example there (e.g. test_files)
|
Hi,
Moved it to the root folder
Added a small validation dataset download so that it works end-to-end, tested it locally |
|
fixed the imports and path issues |
|
looks like all is fixed |
| # see https://github.com/snakers4/silero-models for utils and more examples | ||
|
|
||
| device = torch.device('cpu') # gpu also works, but our models are fast enough for CPU | ||
| model, decoder, utils = torch.hub.load(github='snakers4/silero-models', |
There was a problem hiding this comment.
would this require omegaconf and torchaudio to be installed?
if so, you should add a cell like in https://github.com/pytorch/hub/blob/master/nvidia_deeplearningexamples_waveglow.md#example
which pip installs these extra packages.
that will make the thing instantly run in Google Colab, and is really valuable!
There was a problem hiding this comment.
Hi,
would this require omegaconf and torchaudio to be installed?
Yeah, this would
That is why I needed to include them in your CI environment above
Actually in colab I also needed to install soundfile, as we are using it as a backend for TorchAudio
if so, you should add a cell like in
that will make the thing instantly run in Google Colab, and is really valuable!
yeah, this totally makes sense
by the way, you can see a more extended colab version here
I will add this shortly
snakers4_silero-models_stt.md
Outdated
| read_audio, | ||
| prepare_model_input) = utils # see function signature for details | ||
|
|
||
| torch.hub.download_url_to_file('http://www.openslr.org/resources/83/midlands_english_female.zip', |
There was a problem hiding this comment.
instead of a zip file, extracting it and then running a batch of wav files through, it would be much nicer and more illustrative of downloading a single wav file and processing it through.
like:
torch.hub.download_url_to_file('some download url for speech.wav', dst='speech.wav)
input = prepare_model_input(glob('midlands_english_female/*.wav'), device=device)
output = model(input)
There was a problem hiding this comment.
one of reasons why I am downloading whole validation dataset and running a batch is to demonstrate the models are actually quite fast on CPU as well as on GPU
but I see your point
I will add a default example with one file, but I would nevertheless keep the utils
because for a first time user batching may be an issue, so I would like to solve that
I will add the changes shortly
|
thanks for the great contribution. it should go live sometime today, maybe in a couple of hours. |
|
checked here I guess it is not there yet, looking forward to the release! on a side note, as a small self-funded team we feel extremely proud to have made something worthy of including in this hub |
|
apparently we've moved to a once-per-day update of the site now. it should be live by tomorrow morning. |
|
it is live now |
Please kindly review our Speech-To-Text models