Which code generated this folder (hemmer_increase_experts_trained) in the "ham1000' file

[hechengbo-H]

Hello, I think your work is particularly excellent. I would like to ask what is used to generate this file in the "ham1000' file (hemmer_increase_experts_trained).

[dbarrejon]

Hi,

I just updated the code so that the folder matches the definition in the script.

The folder hemmer_increase_experts_trained is created from the script hemmer_baseline_trained.py in the HAM10000 experiment. Before it was called hemmer_increase_experts_trained_test in the mentioned script. Hope this could help!

[hechengbo-H]

Excuse me, I ran ham10000/validation_expert_model.py and encountered the following problem: 75 epochs...
Traceback (most recent call last):

What I do know for sure is that device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu") is uniform across programs. Because when I annotate "evaluationConvMixer(testloader, "m_convMixer_expert_v2", device=device)" and run "evaluation_MLPMixer(testloader, device=device) "will give you the result.
Here is the result of the evaluation_MLPMixer(testloader, device=device) function:

I am eager for your help. Thank you very much

[hechengbo-H]

Also, there is no such thing as "cuda:1" in the current file. I'm sure it's all in "cuda:0".

[hechengbo-H]

Thanks I've solved the problem because I needed to change "model = nn.DataParallel(model)" to "model = model.to(device)". Before that I was sure I was using "cuda: 0", but I didn't know why it was wrong.