Coder Social home page Coder Social logo

inlite-hrnet's Introduction

single-gpu testing

python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [--out ${RESULT_FILE}] [--eval ${EVAL_METRIC}] [--proc_per_gpu ${NUM_PROC_PER_GPU}] [--gpu_collect] [--tmpdir ${TMPDIR}] [--average_clips ${AVG_TYPE}] [--launcher ${JOB_LAUNCHER}] [--local_rank ${LOCAL_RANK}]

lite-hrnet

python test.py configs/top_down/lite_hrnet/coco/litehrnet_18_coco_256x192.py weights/litehrnet_18_coco_256x192.pth 3 --out test_litehrnet_18_coco_256x192.txt --eval mAP ./tools/dist_test.sh configs/top_down/lite_hrnet/coco/litehrnet_18_coco_256x192.py weights/litehrnet_18_coco_256x192.pth 3 --eval mAP

python -m torch.distributed.launch --nproc_per_node=3 --master_port=29500 test.py configs/top_down/lite_hrnet/coco/litehrnet_18_coco_256x192.py
weights/litehrnet_18_coco_256x192.pth --launcher pytorch --eval mAP

dite-hrnet

python train.py configs/top_down/dite_hrnet/coco/ditehrnet_18_coco_256x192.py

navie_litehrnet_18_coco_256x192_e210_v3.pth

navie-lite-hrnet

python -m torch.distributed.launch --nproc_per_node=3 --master_port=29500 train.py configs/top_down/naive_litehrnet/coco/wider_naive_litehrnet_18_coco_256x192.py --launcher pytorch

my navie-lite-hrnet train

python -m torch.distributed.launch --nproc_per_node=3 --master_port=29500 train.py configs/top_down/dite_hrnet/coco/wider_naive_litehrnetv3_18_coco_256x192.py --launcher pytorch python -m torch.distributed.launch --nproc_per_node=3 --master_port=29500 train.py configs/top_down/dite_hrnet/coco/inlitehrnet_18_coco_256x192.py --launcher pytorch

test

python -m torch.distributed.launch --nproc_per_node=3 --master_port=29500 test.py configs/top_down/dite_hrnet/coco/wider_naive_litehrnetv3_18_coco_256x192.py weights/navie_litehrnet_18_coco_256x192_e210_v3.pth --launcher pytorch python -m torch.distributed.launch --nproc_per_node=3 --master_port=29500 test.py configs/top_down/dite_hrnet/coco/inlitehrnet_18_coco_256x192.py weights/inlitehrnet_18_coco_256x192.pth --launcher pytorch

test 18 - 384x288

python -m torch.distributed.launch --nproc_per_node=3 --master_port=29500 test.py configs/top_down/dite_hrnet/coco/inlitehrnet_18_coco_384x288.py weights/inlitehrnet_18_coco_384x288_e190.pth --launcher pytorch python -m torch.distributed.launch --nproc_per_node=3 --master_port=29500 test.py configs/top_down/dite_hrnet/coco/inlitehrnet_18_coco_256x192.py weights/navie_litehrnet_18_coco_256x192_e210_v3.pth --launcher pytorch

train mpii

python -m torch.distributed.launch --nproc_per_node=3 --master_port=29500 train.py configs/top_down/dite_hrnet/mpii/inlitehrnet_18_mpii_256x256.py --launcher pytorch python -m torch.distributed.launch --nproc_per_node=3 --master_port=29500 test.py configs/top_down/dite_hrnet/mpii/inlitehrnet_18_mpii_256x256.py work_dirs/inlitehrnet_18_mpii_256x256/epoch_220.pth --launcher pytorch

InLite 30 384*288

python -m torch.distributed.launch --nproc_per_node=3 --master_port=29500 train.py configs/top_down/dite_hrnet/coco/inlitehrnet_30_coco_384x288.py --launcher pytorch CUDA_VISIBLE_DEVICES='2' python train.py configs/top_down/dite_hrnet/coco/wider_naive_litehrnetv3_18_coco_256x192.py CUDA_VISIBLE_DEVICES='1' python train.py configs/top_down/dite_hrnet/coco/inlitehrnet_30_coco_384x288.py CUDA_VISIBLE_DEVICES='0' python train.py configs/top_down/dite_hrnet/coco/inlitehrnet_18_coco_384x288.py

InLiteV1（大论文 + LMA

python -m torch.distributed.launch --nproc_per_node=3 --master_port=29500 train.py configs/top_down/dite_hrnet/coco/inlitehrnetv1_18_coco_256x192.py --launcher pytorch python -m torch.distributed.launch --nproc_per_node=3 --master_port=29500 train.py configs/top_down/dite_hrnet/coco/inlitehrnetv1_18_coco_384x288.py --launcher pytorch

InLiteV1(CInlite-HRNet COCO test值)

python -m torch.distributed.launch --nproc_per_node=3 --master_port=29500 test.py configs/top_down/dite_hrnet/coco/inlitehrnetv1_18_coco_256x192.py weights/inlitehrnetv1_18_coco_256x192.pth --launcher pytorch python -m torch.distributed.launch --nproc_per_node=3 --master_port=29500 test.py configs/top_down/dite_hrnet/coco/inlitehrnetv1_18_coco_384x288.py weights/inlitehrnetv1_18_coco_384x288.pth --launcher pytorch

multiple-gpu testing

./tools/dist_test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} ${GPU_NUM} [--out ${RESULT_FILE}] [--eval ${EVAL_METRIC}]
[--proc_per_gpu ${NUM_PROC_PER_GPU}] [--gpu_collect] [--tmpdir ${TMPDIR}] [--average_clips ${AVG_TYPE}]
[--launcher ${JOB_LAUNCHER}] [--local_rank ${LOCAL_RANK}]

conda create -n ziyang_hrnet --clone ziyang_ditehrnet

计算网络复杂度

python summary_network.py configs/top_down/dite_hrnet/coco/inlitehrnet_18_coco_384x288.py python summary_network.py configs/top_down/dite_hrnet/coco/inlitehrnetv1_18_coco_256x192.py python test_speed.py configs/top_down/dite_hrnet/coco/inlitehrnetv1_18_coco_256x192.py python summary_network.py configs/top_down/dite_hrnet/coco/wider_naive_litehrnetv3_18_coco_256x192.py

可视化log

tensorboard --logdir ${WORK_DIR}/${TIMESTAMP}/vis_data conda activate xinru-pet tensorboard --logdir E:\zzy\LearnWorkspace\LearnDeepLearning\Dite-HRNet-main\work_dirs

DSC 创新点，dropout特征图，算出所有通道的注意力，把最小的一半扔掉，然后剩下的进行GhostNet增值

inlite-hrnet's People

Contributors

aomrsou avatar

Watchers

avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.