ValueError: The `target` has to be an integer tensor.

What is the right way to use dataloaders from PyTorch and generators from TensorFlow? 

This gave the above error 
```
history = model.fit(
    inputs=train_loader,
    epochs=100,
    steps_per_epoch=200,
    validation_data=validation_loader,
    shuffle=True,
    callbacks=[eg.callbacks.ModelCheckpoint("models/high-level", save_best_only=True)],
)
```