Somshubra Majumdar
Somshubra Majumdar
You can use this model, which is a chunk aware model - https://huggingface.co/nvidia/stt_en_fastconformer_hybrid_large_streaming_multi
It's a practical limitation. You can either get a ordinary fast conformer in German or a chunk aware Conformer in English. Depends on what your priority is - streaming or...
We don't support export of the SLU model
For transducer, there were always 3 outputs in forward because it's just end encoder forward. During decoding we take those encoder outputs and pass them auto regressively to the decoder...
Transformer engine is not that easy to install, please look at installation instructions in NeMo readme or use container.