Hi, I tried to run the test set by using:
log_preds,y = learn.TTA() probs = np.mean(np.exp(log_preds),0) accuracy_np(probs,y)
But I keep getting the following error:
TypeError: mean() missing 3 required positional argument: “dim”, “keepdim”, “dtype”
I have 4 classes and I’m using a densenet.