Missing key in loading model

[xljh0520]

Hi, thanks for sharing your code. I run the demo

python3 tools/demo.py --config=configs/smpl/tune.py --image_folder=demo_images/ --output_folder=results/ --ckpt data/checkpoint.pt

However, I get this info:

unexpected key in source state_dict: fc.weight, fc.bias

missing keys in source state_dict: layer2.1.bn3.num_batches_tracked, layer3.2.bn2.num_batches_tracked, layer2.2.bn2.num_batches_tracked, layer1.0.bn1.num_batches_tracked, layer4.1.bn3.num_batches_tracked, layer1.2.bn2.num_batches_tracked, layer4.0.bn1.num_batches_tracked, layer1.1.bn1.num_batches_tracked, layer1.2.bn1.num_batches_tracked, layer1.1.bn3.num_batches_tracked, layer4.2.bn3.num_batches_tracked, layer4.1.bn2.num_batches_tracked, layer4.0.bn3.num_batches_tracked, layer2.3.bn1.num_batches_tracked, layer3.5.bn3.num_batches_tracked, layer3.3.bn2.num_batches_tracked, layer4.0.downsample.1.num_batches_tracked, layer4.2.bn2.num_batches_tracked, layer2.0.bn1.num_batches_tracked, layer3.4.bn1.num_batches_tracked, layer3.1.bn3.num_batches_tracked, layer3.2.bn1.num_batches_tracked, layer3.3.bn1.num_batches_tracked, layer2.3.bn2.num_batches_tracked, layer3.5.bn2.num_batches_tracked, layer4.2.bn1.num_batches_tracked, layer3.1.bn1.num_batches_tracked, layer1.2.bn3.num_batches_tracked, layer4.1.bn1.num_batches_tracked, layer3.0.bn1.num_batches_tracked, layer2.0.bn3.num_batches_tracked, layer2.1.bn1.num_batches_tracked, layer2.0.bn2.num_batches_tracked, layer2.3.bn3.num_batches_tracked, layer1.0.bn3.num_batches_tracked, layer3.0.downsample.1.num_batches_tracked, layer3.0.bn2.num_batches_tracked, layer3.2.bn3.num_batches_tracked, layer3.4.bn2.num_batches_tracked, layer2.2.bn1.num_batches_tracked, layer3.0.bn3.num_batches_tracked, layer1.0.bn2.num_batches_tracked, layer3.5.bn1.num_batches_tracked, layer1.0.downsample.1.num_batches_tracked, layer2.2.bn3.num_batches_tracked, layer3.1.bn2.num_batches_tracked, layer4.0.bn2.num_batches_tracked, layer2.0.downsample.1.num_batches_tracked, layer3.3.bn3.num_batches_tracked, layer1.1.bn2.num_batches_tracked, bn1.num_batches_tracked, layer2.1.bn2.num_batches_tracked, layer3.4.bn3.num_batches_tracked

2022-08-19 12:51:02,959 - INFO - load checkpoint from data/checkpoint.pt
2022-08-19 12:51:03,424 - WARNING - missing keys in source state_dict: smpl_head.loss.smpl.J_regressor_extra, smpl_head.smpl.v_template, smpl_head.loss.smpl.faces_tensor, smpl_head.loss.smpl.posedirs, smpl_head.smpl.J_regressor_extra, smpl_head.smpl.parents, smpl_head.loss.smpl.lbs_weights, smpl_head.loss.smpl.parents, smpl_head.smpl.shapedirs, smpl_head.smpl.vertex_joint_selector.extra_joints_idxs, smpl_head.smpl.faces_tensor, smpl_head.smpl.posedirs, smpl_head.smpl.lbs_weights, smpl_head.smpl.J_regressor, smpl_head.loss.smpl.vertex_joint_selector.extra_joints_idxs, smpl_head.loss.smpl.v_template, smpl_head.loss.smpl.shapedirs, smpl_head.loss.smpl.J_regressor

2022-08-19 12:51:03,450 - INFO - resumed epoch 27, iter 90901

I suppose there are many missing keys so the demo will fail. However, I got a reasonable result. I tried as #1 . But I still get the same info.

[JiangWenPL]

I think these missing keys might be a version issue of mmcv or implementations of batchnorm or optimizers. num_batches_tracked should not affect the behaviors during inference.