test failed after train new model using demo

facebookresearch / DensePose

A real-time approach for mapping all human pixels of 2D RGB images to a 3D surface-based model of the body

http://densepose.org

Other

6.98k stars 1.3k forks source link

test failed after train new model using demo #30

Closed waterbroz closed 6 years ago

waterbroz commented 6 years ago

thanks for sharing your works. I try to train the model using the the demo in your GETTING_STARTED.md

python2 tools/train_net.py \
    --cfg configs/DensePose_ResNet50_FPN_single_GPU.yaml \
    OUTPUT_DIR /tmp/detectron-output

How can i test after trainning? I get trouble (KeyError: u'rois'). After checking the code, i found inputs can not get boxes because of the FASTTRCNN:False in yaml file.

vkhalidov commented 6 years ago

You either need to train an end-to-end model (e.g. DensePose_ResNet50_FPN_s1x-e2e) or provide bounding box proposals (see here)