Skip to content

maktu6/garment_segmentation

Repository files navigation

Garment Segmentation

Requirements

Note: please install nightly build gluoncv for less bugs

pip install gluoncv --pre --upgrade

Dataset

iMaterialist (Fashion) 2019 at FGVC6

  • Convert imaterialist to COCO style
python tools/convert_imaterialist_to_coco.py --data_root datasets/imaterialist
  • files structure
datasets
└─imaterialist
    ├── annotations
    │   ├── resize_rle_instances_val.json
    │   ├── rle_instances_train.json
    |   └── rle_instances_val.json
    ├── test
    ├── train
    ├── val
    ├── label_descriptions.json
    ├── sample_submission.csv
    └── train.csv

Training

Note: all using COCO pretrained models, only the categories less than 13 (main apparels) for segmentation.

python train_mask_rcnn.py --save-prefix train_logs/xxx/ --val-interval 1
python train_segmentation.py --model deeplabv3 --dataset imaterialist \
                            --model-zoo deeplab_resnet101_coco --aux \
                            --checkname res101 --epochs 30 --lr 0.001 --ngpus 1 \
                            --workers 2 --batch-size 2 --test-batch-size 2
python train_segmentation.py --model deeplabv3plus --dataset imaterialist \
                             --model-zoo deeplab_plus_xception_coco --aux \
                             --base-size 576 --crop-size 512 \
                             --checkname xception --epochs 30 --lr 0.001 --ngpus 1 \
                             --workers 1 --batch-size 1 --test-batch-size 1 

Reference

Releases

No releases published

Packages

No packages published

Languages