Speech Recognition System & Single Image Super-Resolution
PyTorch, NLP, CV, Hyperparameter tuning, Neural Architecture Search
Instructors: Bhiksha Raj and Rita Singh
Project summary
- Developed with PyTorch an end-to-end encoder-decoder RNN-based speech recognition system for transcript generation.
- Implemented generative single image super-resolution systems such as VAE, SRGAN, and ESRGAN and investigated the effect of different losses on model performance.
Xiangyu Yin
Postdoc @ ANL | AI4science, Physics4ML, scientific discovery acceleration & automation