Closed zideliu closed 10 months ago
Thank you for your open source! There is a small problem here. I read the code you wrote and found that the pre-training model was not used during training, right?
Hi, thanks for your interests. Yes, we train our network from the scratch, i.e., we donot use a Palette-style pre-trained model.
Thank you for your open source! There is a small problem here. I read the code you wrote and found that the pre-training model was not used during training, right?