Different results with multi-GPUs

thuml / CDAN

Code release for "Conditional Adversarial Domain Adaptation" (NIPS 2018)

399 stars 92 forks source link

Open Marsrocky opened 4 years ago

Marsrocky commented 4 years ago

I tested the results using one GPU and multi-GPUs in the same server on Office-31. The results are different.

For the CDAN+E on A->W task: One GPU: around 75% Multi-GPUs (>=2): 92%

I am still investigating the reason.

BX-xb commented 3 years ago

Hi, I have a question usps_train_image = np.load("../data/usps/train_image.npy") How can I get usps dataset?