This is the repository for the paper "N for Parameter: Efficient Neural Audio Super resolution with GAN"
To train own model with fixed loss weights run the file model.py
To train own model with loss balancer from ENCODEC paper run the file load_balanced_model.py
To run test script run the file test_gpu.py

• Update lines 34, 55 and 60 in test_gpu.py to change input sample rate
• For details of settings see the repository https://github.com/haoheliu/ssr_eval/tree/main

We have reused code from the following repositories

1. https://github.com/haoheliu/ssr_eval/tree/main
2. https://github.com/ZhikangNiu/encodec-pytorch
3. https://github.com/kaiidams/soundstream-pytorch