Error during training on 12 view dataset
piseabhijeet opened this issue · comments
Hi @weixmath
I am trying to train the model on a custom dataset with 4 classes and 12 views. I am getting the below error:
Stage 1 training goes fine but the error comes in stage 2 where my batch size is:
train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=20, shuffle=False, num_workers=0)
val_loader = torch.utils.data.DataLoader(val_dataset, batch_size=15, shuffle=False, num_workers=0)
I have changed the classes and the output tensor in the necessary places too . I am using Tesla V100 32GB GPU for training.
Could you please help?
Hi @weixmath
Can you explain how the numbers are derived from line 62 below - Trainer.py:
I am guessing it is calculated for 20 views and 40 classes. Can we derive for 12 views and 4 classes?
Thanks
@piseabhijeet
num_train_files: 12520
num_val_files: 2000
Traceback (most recent call last):
File "/view-GCN-master/train.py", line 66, in
trainer.train(30)
File "\view-GCN-master\tools\Trainer.py", line 52, in train
for i, data in enumerate(self.train_loader):
File "D:\soft_ware\Anaconda3\envs\mvcnn\lib\site-packages\torch\utils\data\dataloader.py", line 819, in next
return self._process_data(data)
File "D:\soft_ware\Anaconda3\envs\mvcnn\lib\site-packages\torch\utils\data\dataloader.py", line 846, in _process_data
data.reraise()
File "D:\soft_ware\Anaconda3\envs\mvcnn\lib\site-packages\torch_utils.py", line 369, in reraise
raise self.exc_type(msg)
ValueError: Caught ValueError in DataLoader worker process 0.
Original Traceback (most recent call last):
File "D:\soft_ware\Anaconda3\envs\mvcnn\lib\site-packages\torch\utils\data_utils\worker.py", line 178, in _worker_loop
data = fetcher.fetch(index)
File "D:\soft_ware\Anaconda3\envs\mvcnn\lib\site-packages\torch\utils\data_utils\fetch.py", line 44, in fetch
data = [self.dataset[idx] for idx in possibly_batched_index]
File "D:\soft_ware\Anaconda3\envs\mvcnn\lib\site-packages\torch\utils\data_utils\fetch.py", line 44, in
data = [self.dataset[idx] for idx in possibly_batched_index]
File "\view-GCN-master\tools\ImgDataset.py", line 112, in getitem
class_id = self.classnames.index(class_name)
ValueError: 'modelnet40v2png_ori4' is not in list
hi! How to resolve these two errors at run time?
thanks
I had tried this repo almost an year ago. Unfortunately at this point of time, I do not have access to the resource nor the dataset which I had worked on. Apologies for the same.