search

jimmyyhwu / pose-interpreter-networks

Real-time robotic object pose estimation with deep learning
MIT License
122 stars 27 forks source link

steps to train model again with another DRN #28

Closed meghalD closed 4 years ago

meghalD commented 4 years ago

Hello, I want to train another drn model on the same dataset. Please guide me through steps for end to end evaluation. Also, I can see DRNSeg being called during training. But where is drn22 being called by DRNSeg ?

jimmyyhwu commented 4 years ago

I believe the segmentation README should guide you through the steps.

meghalD commented 4 years ago

I tried running python train.py config/drn_d_22_OilChange.yml. I throws an error as follows:

log_dir: logs/2019-10-31_03-23-27.170915_drn_d_22_OilChange checkpoint_dir: checkpoints/2019-10-31_03-23-27.170915_drn_d_22_OilChange loading annotations into memory... Done (t=0.46s) creating index... index created! loading annotations into memory... Done (t=0.07s) creating index... index created! Traceback (most recent call last): File "train.py", line 238, in main(cfg) File "train.py", line 202, in main train_batch_time, train_data_time, train_loss = train(train_loader, model, criterion, optimizer, epoch) File "train.py", line 45, in train for i, (input, target) in enumerate(train_loader): File "/home/meghal/python_iros/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 336, in next return self._process_next_batch(batch) File "/home/meghal/python_iros/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 357, in _process_next_batch raise batch.exc_type(batch.exc_msg) AttributeError: Traceback (most recent call last): File "/home/meghal/python_iros/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 106, in _worker_loop samples = collate_fn([dataset[i] for i in batch_indices]) File "/home/meghal/python_iros/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 106, in samples = collate_fn([dataset[i] for i in batch_indices]) File "/home/meghal/pose-interpreter-networks/segmentation/datasets.py", line 42, in getitem image, label = self.transforms([image, label]) File "/home/meghal/pose-interpreter-networks/segmentation/transforms.py", line 211, in call return [self._transform(self.transforms[i], d, i) for i, d in enumerate(data)] File "/home/meghal/pose-interpreter-networks/segmentation/transforms.py", line 211, in return [self._transform(self.transforms[i], d, i) for i, d in enumerate(data)] File "/home/meghal/pose-interpreter-networks/segmentation/transforms.py", line 207, in _transform data = transform(data) File "/home/meghal/pose-interpreter-networks/segmentation/transforms.py", line 173, in call return adjust_gamma(data, gamma=value) File "/home/meghal/python_iros/lib/python3.6/site-packages/skimage/exposure/exposure.py", line 419, in adjust_gamma return out.astype(dtype) AttributeError: 'Tensor' object has no attribute 'astype'

meghalD commented 4 years ago

Googling for the error shows similar problems with other attributes. As far we have checked, none of them serve our purpose. Any pointers will be highly appreciated.

jimmyyhwu commented 4 years ago

Try downgrading scikit-image, I think some of the libraries have changed in the past year, breaking the code.

pip install scikit-image==0.14.2
meghalD commented 4 years ago

Try downgrading scikit-image, I think some of the libraries have changed in the past year, breaking the code.

pip install scikit-image==0.14.2

It worked after downgrading to 0.14.2. Thanks.