boostcampaitech3 / level2-semantic-segmentation-level2-cv-09

level2-semantic-segmentation-level2-cv-09 created by GitHub Classroom

2 stars 4 forks source link

[Model] MMSegmentation BEiT setting #54

Open updaun opened 2 years ago

updaun commented 2 years ago

BEiT config file setting
train and inference

updaun commented 2 years ago

custom_mmcv install

git clone https://github.com/microsoft/unilm.git

move directory custom_mmcv

unilm/beit/semantic_segmentation/mmcv_custom-> mmsegmentation/mmcv_custom

apex install

git clone https://github.com/NVIDIA/apex cd apex python setup.py install

beit pretrained model download

Download checkpoints that are self-supervised pretrained and then intermediate fine-tuned on ImageNet-22k (recommended):

BEiT-base: beit_base_patch16_224_pt22k_ft22k
BEiT-large: beit_large_patch16_224_pt22k_ft22k

mmcv-full install

pip install --ignore-installed mmcv-full==1.3.0 -f https://download.openmmlab.com/mmcv/dist/cu110/torch1.7.0/index.html

학습 중 에러 발생 시 (버전업 1.3.13)

pip install --ignore-installed mmcv-full==1.3.13 -f https://download.openmmlab.com/mmcv/dist/cu110/torch1.7.0/index.html

64_BEiT download

https://github.com/boostcampaitech2/semantic-segmentation-level2-cv-16/tree/main/release/segmentation_model/BEiT/64_BEiT 다운로드 후 위치를 찾아 해당 소스로 변경

backbone/beit.py ->/opt/ml/input/code/mmsegmentation/mmseg/models/backbones/beit.py
config 5files
tools 4files

jeongjae96 commented 2 years ago

SwinL 학습 끝나고 도전해보겠습니다!