Simpler way to get predicted results in batches after training

1

Hi, I am trying to get predicted results after model training. Since the input data is too large to fit into memory, I would need to divide the data into batches for evaluation. I know it is possible to do the following in pytorch:

with torch.no_grad():
    result_list = []
    for batch in dataloader:
        result = net(batch)
        result_list.append(result)
# then concatenate list to get predicted results

I am just wondering if there are simpler ways to do this, something like mode.predict(input, batch_size=batch_size) in Keras where batch_size option automatically takes care of the splitting and recombination process?

Thank you!

2

For a higher-level API you could take a look at e.g. Ignite, Catalyst, and Lightning, which I guess are providing these kind of APIs.