Using module to get result

[beeroom]

Hi @magehrig
I noticed that there are a lot of integrated writing methods in your code， but unfortunately I'm not very familiar with it. For example, for data loader, you used the code:

data_module = fetch_data_module(config=config)

for model:

 module = fetch_model_module(config=config)
 module = module.load_from_checkpoint(str(ckpt_path), **{'full_config': config})

Could you let me know if there is a simple way to extract some data from data_module and then input it into the module to get a result?

[magehrig]

Hi @beeroom

I can only specific questions about the project. For general questions regarding APIs, you have to check out the Pytorch documentation.

[beeroom]

OK, I got it. I know how to do it already.
But I still have an another question: have you provided the code for testing the inference time just as you provided in Chapter 4?
I run the validation.py but don't see there is a metric about the inference time.

[magehrig]

this code is for reproducing the quantitative detection performance, not the inference time. It should be relatively straightforward to adapt the code if you want to test inference time.

[beeroom]

OK. I will try it!