Step 6: Connecting Hub Datasets to ML Frameworks

You can connect Hub Datasets to popular ML frameworks such as PyTorch and TensorFlow using minimal boilerplate code, and Hub takes care of the parallel processing!
PyTorch
You can train a model by creating a PyTorch DataLoader from a Hub Dataset using ds.pytorch() .
import hub
from torch.utils.data import DataLoader
​
ds = hub.dataset('./dataset_path') # Hub Dataset
dataloader= ds.pytorch(batch_size = 16, num_workers = 2) #PyTorch Dataloader
​
for data in dataloader:
    print(data)
    # Training Loop
TensorFlow - Coming Soon
Similarly, you can convert a Hub Dataset to a TensorFlow Dataset via the tf.Data API. 
ds # Hub Dataset object, to be used for training
ds = ds.tensorflow() # A TensorFlow Dataset

Step 5: Using Activeloop Storage

Step 7: Parallel Computing

Last modified 1yr ago

Copy link

Outline

PyTorch

TensorFlow - Coming Soon