search

dougsm / ggcnn

Generative Grasping CNN from "Closing the Loop for Robotic Grasping: A Real-time, Generative Grasp Synthesis Approach" (RSS 2018)
BSD 3-Clause "New" or "Revised" License
484 stars 138 forks source link

AttributeError #57

Open normajeane-m opened 9 months ago

normajeane-m commented 9 months ago

I appreciated your amazing research.

After the pcd files were converted to depth images (using function generate_cornell_depth), I got some errors when I tried to train the network with Cornell dataset.

python3 train_ggcnn.py --description training --network ggcnn2 --dataset cornell --dataset-path cornell/

AttributeError: Caught AttributeError in DataLoader worker process 0
AttributeError: Couldn't find function center in BoundingBoxes or BoundingBox

There was same result with evaluation code with pre-trained model.

python3 eval_ggcnn.py --network ggcnn_weights_cornell/ggcnn_epoch_23_cornell --dataset cornell --dataset-path cornell --iou-eval

AttributeError: Caught AttributeError in DataLoader worker process 0
AttributeError: Couldn't find function center in BoundingBoxes or BoundingBox

But, I tested the task to loading the ggcnn using pre-trained model.

from models.ggcnn import GGCNN model = GGCNN() model.load_state_dict(torch.load('ggcnn_weights_cornell/ggcnn_epoch_23_cornell_statedict.pt'))

All keys matched successfully

I would like to thank you once again.

kalyanksk commented 9 months ago

I also faced the issue and changing numpy version worked for me!

pip uninstall numpy
pip install numpy==1.22
t-272 commented 2 months ago

I also met this two questions.