MaX-DeepLab

Unofficial implementation of MaX-DeepLab for Instance Segmentation: https://arxiv.org/abs/2012.00759v1.

Status

Only the MaX-DeepLab-S architecture is putatively implemented. Primarily, this code is intended as a reference; I can't make any guarantees that it will reproduce the results of the paper.

• [x] Axial Attention block
• [x] Dual Path Transformer block
• [x] MaX-DeepLab-S architecture
• [x] Hungarian Matcher
• [x] PQ-style loss
• [x] Auxiliary losses (Instance discrimination, Mask-ID cross-entropy, Semantic Segmentation)
• [x] Coco Panoptic Dataset
• [x] Simple inference

Usage

See example.ipynb.