-
Hi. Im want to developing a streaming interface for my trainned Conformer Transducer ASR model. I have only see tutorials about streaming for CTC model here:
https://github.com/NVIDIA/NeMo/blob/stab…
-
i saw beta version of transducer loss relies on code from Alignment Restricted Streaming Recurrent Neural Network Transducer , maybe it's has parameters the left and right buffers bl/lr, how can i use…
-
I followed same steps discussed here https://github.com/k2-fsa/k2/issues/952
I created `activate.sh` with the following:
export CUDA_HOME=/cm/shared/apps/cuda11.0/toolkit/11.0.3
export PATH=$C…
-
I try perform steps from https://colab.research.google.com/github/NVIDIA/NeMo/blob/stable/tutorials/asr/ASR_with_Transducers.ipynb witn my own dataset.
After 10% of first epoch i see in tensorboar…
-
So, I have implemented the vanilla RNN-T (Transducer) model in Tensorflow and also a `greedy_decode()` function similar to
https://github.com/espnet/espnet/blob/88bb9dba83636c4bf12f10a870a7d8c5f951…
-
Hi, thanks for this great package! I'm exploring using this approach for a real-time application. It took me a moment to realize how to use this with values that are acquired in a streaming fashion, s…
-
torchaudio is targeting speech recognition as full audio application ([internal](https://docs.google.com/document/d/1mygzn6vf6Jl4e_0fzJSciRGRmb8dj8NwibwBG-tyUj8/edit?ts=5ef3580b#bookmark=id.7evu2kt47e…
-
There are places in the results pipeline that we need to alter individual cells in the results. Remappings are an example of this and formatting dates differently (https://github.com/metabase/metabase…
-
Hi,
I tried to infer on a stream_transducer-tflite model (refer to 'examples/demonstration/streaming_tflite_conformer.py'), then some error ocurred (looks like [issue](https://github.com/TensorSpeech…
-
在Transducer类中的perform_greedy方法里,循环中的pred_net用到了lstm进行解码,每次都会对所有decoded的完整序列走一遍lstm,实际只需要解码最后一步来判断是否输出0或者blank即可,在长序列预测下能节省大量时间![image](https://user-images.githubusercontent.com/44560439/129701060-46bd…