A speech de-noiser utilizing a convolutional AutoEncoder

Image denoiser to reduce noise in transmitted documents.

A PyTorch implementation of UNet-based image denoising. Encoder-decoder with skip connections: encoder (3→64→128→256), bottleneck (512), decoder (512→256→128→64) via transposed conv. Outputs 3-channel clean images. For RGB denoising, low-light enhancement.

Autoencoder denoiser on MNIST Dataset - testing ClearML integration