Flow to train the model.

[talk2sohail]

Hi there,

I am trying to train the CNN model but I am not able to prepare the dataset. Can you please help with the flow or steps to be specific in order to achieve the above said purpose. I urge u to please let me know if it can done. My aim is to score protein-protein interaction.

Steps I have done so far:

1. Tried running the train.py with pointclouds.h5 file but I am getting the following error:
   cmd : python3 scripts/train.py --data_path pointclouds.h5 --optimizer SGD
   output :
   Traceback (most recent call last): File "scripts/train.py", line 64, in <module> dataloader = data.DataLoader(dataset, batch_size=arg.batch_size, shuffle=True, num_workers=1) File "/usr/local/lib/python3.7/dist-packages/torch/utils/data/dataloader.py", line 266, in __init__ sampler = RandomSampler(dataset, generator=generator) # type: ignore File "/usr/local/lib/python3.7/dist-packages/torch/utils/data/sampler.py", line 104, in __init__ "value, but got num_samples={}".format(self.num_samples)) ValueError: num_samples should be a positive integer value, but got num_samples=0