Pre-trained models take 2 - tracker

On each of the datasets provided, we must train a Deepspeech model and document exactly how to reproduce these results! The overall architecture is encompassed in this command:

python train.py  --rnn_type gru --hidden_size 800 --hidden_layers 5 --checkpoint --visdom --train_manifest /path/to/train_manifest.csv --val_manifest /path/to/val_manifest.csv --epochs 100 --num_workers $(nproc) --cuda

In the above command you must replace the manifests paths with the correct paths to the dataset. A few notes:

• Train till convergence (should get a nice smooth training curve hopefully!)
• For smaller datasets, you may need to reduce the learning rate annealing by adding the flag --learning anneal and setting it to a smaller value, like 1.01. For larger datasets, the default is fine (up to around 4.5k hours from internal testing on the deepspeech.torch version)

If you use augmentation or noise injection give as much details as possible for me to wrap up it all nicely.

A release will be cut from the DeepSpeech package that will have the models, and a reference to the latest release added to the README to find latest models!

Progress tracker for datasets:

• AN4
• TEDLium
• LibriSpeech

Let me know if you plan on working on running any of these, and I’ll update the ticket with details!

[oplatek]

Hi guys, as above is there any unidirectional model? If yes is there a streaming interface for it?

Hints appreciated in what needs to be done to have both a) the model b) the streaming decoding interface

[yyang29]

Hi guys, just to confirm, is there any unidirectional model released yet?