kywind / self-corr-pose Goto Github PK
View Code? Open in Web Editor NEWSelf-Supervised Geometric Correspondence for Category-Level 6D Object Pose Estimation in the Wild, ICLR 2023
Self-Supervised Geometric Correspondence for Category-Level 6D Object Pose Estimation in the Wild, ICLR 2023
gdown --id 1MOeWKoHoBK9GH6jJ-BZPvD9rj9xQdWux
/home/yuanweizhong/anaconda3/envs/PoseReloc/lib/python3.9/site-packages/gdown/cli.py:126: FutureWarning: Option --id
was deprecated in version 4.3.1 and will be removed in 5.0. You don't need to pass it anymore to use a file ID.
warnings.warn(
Access denied with the following error:
Cannot retrieve the public link of the file. You may need to change
the permission to 'Anyone with the link', or have had many accesses.
You may still be able to access the file from the browser:
https://drive.google.com/uc?id=1MOeWKoHoBK9GH6jJ-BZPvD9rj9xQdWux
Can you share the results of each class on the REAL275 and Wild6D (Related to Table 1 & Table 2 in the main paper)?
Thanks for sharing the code.
I have an issue related to backpropagation during training. Do you have any advice related to this?
using distributed training...
world size: 1, local rank: 0
total number of videos: 190
total number of valid frames: 51854
per gpu batch size: 32
number of gpus: 1
number of iterations per gpu 20000
found vert_params: mesh.mean_v, shape: [995, 3], requires_grad: True
*found unknown params: mesh.faces, shape: [1986, 3], requires_grad: False
*found unknown params: mesh.symm_rots, shape: [2, 3, 3], requires_grad: False
found backbone_params: encoder.backbone.resnet.conv1.weight, shape: [64, 3, 7, 7], requires_grad: True
found backbone_params: encoder.backbone.resnet.bn1.weight, shape: [64], requires_grad: False
found backbone_params: encoder.backbone.resnet.bn1.bias, shape: [64], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer1.0.conv1.weight, shape: [64, 64, 3, 3], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer1.0.bn1.weight, shape: [64], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer1.0.bn1.bias, shape: [64], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer1.0.conv2.weight, shape: [64, 64, 3, 3], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer1.0.bn2.weight, shape: [64], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer1.0.bn2.bias, shape: [64], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer1.1.conv1.weight, shape: [64, 64, 3, 3], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer1.1.bn1.weight, shape: [64], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer1.1.bn1.bias, shape: [64], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer1.1.conv2.weight, shape: [64, 64, 3, 3], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer1.1.bn2.weight, shape: [64], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer1.1.bn2.bias, shape: [64], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer2.0.conv1.weight, shape: [128, 64, 3, 3], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer2.0.bn1.weight, shape: [128], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer2.0.bn1.bias, shape: [128], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer2.0.conv2.weight, shape: [128, 128, 3, 3], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer2.0.bn2.weight, shape: [128], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer2.0.bn2.bias, shape: [128], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer2.0.downsample.0.weight, shape: [128, 64, 1, 1], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer2.0.downsample.1.weight, shape: [128], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer2.0.downsample.1.bias, shape: [128], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer2.1.conv1.weight, shape: [128, 128, 3, 3], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer2.1.bn1.weight, shape: [128], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer2.1.bn1.bias, shape: [128], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer2.1.conv2.weight, shape: [128, 128, 3, 3], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer2.1.bn2.weight, shape: [128], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer2.1.bn2.bias, shape: [128], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer3.0.conv1.weight, shape: [256, 128, 3, 3], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer3.0.bn1.weight, shape: [256], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer3.0.bn1.bias, shape: [256], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer3.0.conv2.weight, shape: [256, 256, 3, 3], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer3.0.bn2.weight, shape: [256], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer3.0.bn2.bias, shape: [256], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer3.0.downsample.0.weight, shape: [256, 128, 1, 1], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer3.0.downsample.1.weight, shape: [256], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer3.0.downsample.1.bias, shape: [256], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer3.1.conv1.weight, shape: [256, 256, 3, 3], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer3.1.bn1.weight, shape: [256], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer3.1.bn1.bias, shape: [256], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer3.1.conv2.weight, shape: [256, 256, 3, 3], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer3.1.bn2.weight, shape: [256], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer3.1.bn2.bias, shape: [256], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer4.0.conv1.weight, shape: [512, 256, 3, 3], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer4.0.bn1.weight, shape: [512], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer4.0.bn1.bias, shape: [512], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer4.0.conv2.weight, shape: [512, 512, 3, 3], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer4.0.bn2.weight, shape: [512], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer4.0.bn2.bias, shape: [512], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer4.0.downsample.0.weight, shape: [512, 256, 1, 1], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer4.0.downsample.1.weight, shape: [512], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer4.0.downsample.1.bias, shape: [512], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer4.1.conv1.weight, shape: [512, 512, 3, 3], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer4.1.bn1.weight, shape: [512], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer4.1.bn1.bias, shape: [512], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer4.1.conv2.weight, shape: [512, 512, 3, 3], requires_grad: True
found backbone_params: encoder.backbone.resnet.layer4.1.bn2.weight, shape: [512], requires_grad: False
found backbone_params: encoder.backbone.resnet.layer4.1.bn2.bias, shape: [512], requires_grad: False
found feat_params: encoder.featnet.upconv5.cbr_unit.0.weight, shape: [256, 512, 3, 3], requires_grad: True
found feat_params: encoder.featnet.upconv5.cbr_unit.0.bias, shape: [256], requires_grad: True
found feat_params: encoder.featnet.iconv4.cbr_unit.0.weight, shape: [256, 512, 3, 3], requires_grad: True
found feat_params: encoder.featnet.iconv4.cbr_unit.0.bias, shape: [256], requires_grad: True
found feat_params: encoder.featnet.upconv4.cbr_unit.0.weight, shape: [128, 256, 3, 3], requires_grad: True
found feat_params: encoder.featnet.upconv4.cbr_unit.0.bias, shape: [128], requires_grad: True
found feat_params: encoder.featnet.iconv3.cbr_unit.0.weight, shape: [128, 256, 3, 3], requires_grad: True
found feat_params: encoder.featnet.iconv3.cbr_unit.0.bias, shape: [128], requires_grad: True
found feat_params: encoder.featnet.upconv3.cbr_unit.0.weight, shape: [64, 128, 3, 3], requires_grad: True
found feat_params: encoder.featnet.upconv3.cbr_unit.0.bias, shape: [64], requires_grad: True
found feat_params: encoder.featnet.iconv2.cbr_unit.0.weight, shape: [64, 128, 3, 3], requires_grad: True
found feat_params: encoder.featnet.iconv2.cbr_unit.0.bias, shape: [64], requires_grad: True
found feat_params: encoder.featnet.proj.weight, shape: [64, 64, 1, 1], requires_grad: True
found feat_params: encoder.featnet.proj.bias, shape: [64], requires_grad: True
found feat_params: encoder.featnet_mesh.stn.conv1.weight, shape: [128, 3, 1], requires_grad: True
found feat_params: encoder.featnet_mesh.stn.conv1.bias, shape: [128], requires_grad: True
found feat_params: encoder.featnet_mesh.stn.fc.weight, shape: [9, 128], requires_grad: True
found feat_params: encoder.featnet_mesh.stn.fc.bias, shape: [9], requires_grad: True
found feat_params: encoder.featnet_mesh.conv1.weight, shape: [64, 3, 1], requires_grad: True
found feat_params: encoder.featnet_mesh.conv1.bias, shape: [64], requires_grad: True
found shape_params: encoder.shape_code_predictor.weight, shape: [64, 512], requires_grad: True
found shape_params: encoder.shape_code_predictor.bias, shape: [64], requires_grad: True
found shape_params: encoder.shape_predictor.shapenerf.layer1.weight, shape: [256, 67], requires_grad: True
found shape_params: encoder.shape_predictor.shapenerf.layer1.bias, shape: [256], requires_grad: True
found shape_params: encoder.shape_predictor.shapenerf.layers_xyz.0.weight, shape: [256, 256], requires_grad: True
found shape_params: encoder.shape_predictor.shapenerf.layers_xyz.0.bias, shape: [256], requires_grad: True
found shape_params: encoder.shape_predictor.shapenerf.layers_dir.0.weight, shape: [128, 256], requires_grad: True
found shape_params: encoder.shape_predictor.shapenerf.layers_dir.0.bias, shape: [128], requires_grad: True
found shape_params: encoder.shape_predictor.shapenerf.fc_alpha.weight, shape: [1, 256], requires_grad: True
found shape_params: encoder.shape_predictor.shapenerf.fc_alpha.bias, shape: [1], requires_grad: True
found shape_params: encoder.shape_predictor.shapenerf.fc_rgb.weight, shape: [3, 128], requires_grad: True
found shape_params: encoder.shape_predictor.shapenerf.fc_rgb.bias, shape: [3], requires_grad: True
found shape_params: encoder.shape_predictor.shapenerf.fc_feat.weight, shape: [256, 256], requires_grad: True
found shape_params: encoder.shape_predictor.shapenerf.fc_feat.bias, shape: [256], requires_grad: True
found cam_params: encoder.pose_predictor.x_offset, shape: [1, 3], requires_grad: False
found cam_params: encoder.pose_predictor.y_offset, shape: [1, 3], requires_grad: False
found cam_params: encoder.pose_predictor.rot_pred_layer.0.0.0.weight, shape: [128, 512], requires_grad: True
found cam_params: encoder.pose_predictor.rot_pred_layer.0.0.0.bias, shape: [128], requires_grad: True
found cam_params: encoder.pose_predictor.rot_pred_layer.0.1.0.weight, shape: [128, 128], requires_grad: True
found cam_params: encoder.pose_predictor.rot_pred_layer.0.1.0.bias, shape: [128], requires_grad: True
found cam_params: encoder.pose_predictor.rot_pred_layer.0.2.0.weight, shape: [128, 128], requires_grad: True
found cam_params: encoder.pose_predictor.rot_pred_layer.0.2.0.bias, shape: [128], requires_grad: True
found cam_params: encoder.pose_predictor.rot_pred_layer.1.weight, shape: [6, 128], requires_grad: True
found cam_params: encoder.pose_predictor.rot_pred_layer.1.bias, shape: [6], requires_grad: True
found cam_params: encoder.pose_predictor.trans_pred_layer.weight, shape: [3, 512], requires_grad: True
found cam_params: encoder.pose_predictor.trans_pred_layer.bias, shape: [3], requires_grad: True
Traceback (most recent call last):
File "/home/user/anaconda3/envs/pytorch3d/lib/python3.9/site-packages/absl/app.py", line 308, in run
_run_main(main, args)
File "/home/user/anaconda3/envs/pytorch3d/lib/python3.9/site-packages/absl/app.py", line 254, in _run_main
sys.exit(main(argv))
File "train.py", line 43, in main
trainer.train()
File "/ssd1/workspace/self-corr-pose/model/trainer.py", line 122, in train
total_loss.mean().backward()
File "/home/user/anaconda3/envs/pytorch3d/lib/python3.9/site-packages/torch/_tensor.py", line 487, in backward
torch.autograd.backward(
File "/home/user/anaconda3/envs/pytorch3d/lib/python3.9/site-packages/torch/autograd/init.py", line 197, in backward
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass
RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatTensor [32, 995, 3]], which is output 0 of AsStridedBackward0, is at version 1; expected version 0 instead. Hint: enable anomaly detection to find the operation that failed to compute its gradient, with torch.autograd.set_detect_anomaly(True).
python-BaseException
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.