Best decoder?

[OleguerCanal]

❓ Questions & Help

Hi guys! First of all thank you so much for such an amazing repo :)

I'd like to know if you have some insights on which decoder architecture works best for end2end training for medium-hard audio. Imagine that model size and available of data are not a problem. Have you done some tests or know of some paper comparing them?

[virgile-blg]

Hi @OleguerCanal,
Did you figure out which encoder - decoder pair was the most successful regarding your experiments ?

I am training Contextnet & Conformer encoders together with transducer decoders (not converging at all) and LSTM decoder: lstms are converging but the output predictions seems not perfectly aligned though it outputs correct words (some words are keep being repeated)

I am training on Librispeech for now.

Best