See installation instructions.
See Preparing Datasets for MaskFormer.
More Utilization: See Getting Started with MaskFormer.
Swin-Tiny
python ./train_net.py \
--resume --num-gpus 2 --dist-url auto \
--config-file configs/ade20k-150/swin/vw/vw_maskformer_swin_tiny_bs16_160k.yaml \
OUTPUT_DIR path/to/tiny TEST.EVAL_PERIOD 10000 MODEL.MASK_FORMER.SIZE_DIVISIBILITY 64
Swin-Small
python ./train_net.py \
--resume --num-gpus 4 --dist-url auto \
--config-file configs/ade20k-150/swin/vw/vw_maskformer_swin_small_bs16_160k.yaml \
OUTPUT_DIR path/to/small TEST.EVAL_PERIOD 10000 MODEL.MASK_FORMER.SIZE_DIVISIBILITY 64
Swin-Base
python ./train_net.py \
--resume --num-gpus 8 --dist-url auto \
--config-file configs/ade20k-150/swin/vw/vw_maskformer_swin_base_IN21k_384_bs16_160k_res640.yaml \
OUTPUT_DIR path/to/base TEST.EVAL_PERIOD 10000 MODEL.MASK_FORMER.SIZE_DIVISIBILITY 64
Swin-Large
python ./train_net.py \
--resume --num-gpus 16 --dist-url auto \
--config-file configs/ade20k-150/swin/vw/vw_maskformer_swin_large_IN21k_384_bs16_160k_res640.yaml \
OUTPUT_DIR path/to/large TEST.EVAL_PERIOD 10000 MODEL.MASK_FORMER.SIZE_DIVISIBILITY 64
cd MaskFormer
python ./train_net.py \
--eval-only --num-gpus NGPUS --dist-url auto \
--config-file path/to/config \
MODEL.WEIGHTS path/to/weight TEST.AUG.ENABLED True MODEL.MASK_FORMER.SIZE_DIVISIBILITY 64
| Name | Backbone | crop size |
lr sched |
mIoU | mIoU (ms+flip) |
download |
|---|---|---|---|---|---|---|
| VW-MaskFormer | Swin-T | 512x512 | 160k | 47.4 | 49.0 | model |
| VW-MaskFormer | Swin-S | 512x512 | 160k | 50.5 | 52.7 | model |
| VW-MaskFormer | Swin-B | 640x640 | 160k | 53.8 | 54.6 | model |
| VW-MaskFormer | Swin-L | 640x640 | 160k | 55.3 | 56.5 | model |
@inproceedings{yan2023multi,
title={Multi-Scale Representations by Varing Window Attention for Semantic Segmentation},
author={Yan, Haotian and Wu, Ming and Zhang, Chuang},
booktitle={The Twelfth International Conference on Learning Representations},
year={2023}
}