Dataset: LibriSpeech Training using train-100h
- Using TensorFlow CudnnLSTM to train LSTMs, using warpctc-tensorflow to do ctc_loss calculation(optional).
- Simplify model to enable high performance inference on Jetson TX1
TensorFlow Speech Recognition using LSTMs and ConvNets to realize high performance inference on NVIDIA Jetson (TX1)
Dataset: LibriSpeech Training using train-100h
TensorFlow Speech Recognition using LSTMs and ConvNets to realize high performance inference on NVIDIA Jetson (TX1)
MIT License