Eman22S/Amharic-End-to-End

# End-to-end Automatic Speech Recognition Systems - PyTorch Implementation This is an open source project (formerly named **Listen, Attend and Spell - PyTorch Implementation**) for end-to-end ASR implemented with Pytorch, the well known deep learning toolkit. The end-to-end ASR was based on Listen, Attend and Spell 1 . Multiple techniques proposed recently were also implemented, serving as additional plug-ins for better performance. For the list of techniques implemented, please refer to the highlights, configuration and references. Feel free to use/modify them, any bug report or improvement suggestion will be appreciated. If you have any questions, please contact r07922013[AT]ntu.edu.tw. If you find this project helpful for your research, please do consider to cite my paper, thanks! ## Highlights - Acoustic feature extraction - Purepython extraction using librosa as the backend - One-click execution scripts (currently supporting TIMT & LibriSpeech) - Phoneme/character/subword 2 /word embedding for text encoding - End-to-end ASR - Seq2seq ASR with different types of encoder/attention 3 - CTC-based ASR 4 , which can also be hybrid 5 with the former - *yaml*-styled model construction and hyper parameters setting - Training process visualization with TensorBoard, including attention alignment - Speech recognition (decoding) - Beam search decoding - RNN language model training and joint decoding for ASR 6 - Joint CTC-attention based decoding 6 *You may checkout some example log files with TensorBoard by downloading them from `log/log_url.txt`* ## Requirements - Python 3 - Computing power (high-end GPU) and memory space (both RAM/GPU's RAM) is **extremely important** if you'd like to train your own model. - Required packages and their use are listed here. ## Instructions ***Before you start, make sure all the packages required were installed correctly*** ### Step 0. Preprocessing - Acoustic Feature Extraction & Text Encoding Preprocessing scripts ar …