epns

I tried SwinTransformer with partialfc, the 'embed_dim' of swintransformer is set to 96 and embedding_size of partialfc is set to 512. The img_size is 224*224.

pytorch=1.2.0 torchvision=0.4.0 cuda=10.0 output_inds = torch.from_numpy(output_inds).to(self.opt.device) RuntimeError: sizes must be non-negative

Do you use gradcam (I cannot visualize heatmap by this way) or have a special function

default_dataset_info = { 'evodet':{'default_resolution': [512,512], 'num_classes': 80, 'mean': [0.408, 0.447, 0.470], 'std': [0.289, 0.274, 0.278], 'dataset': 'coco'}, 'ctdet': {'default_resolution': [512, 512], 'num_classes': 80, 'mean': [0.408, 0.447, 0.470], 'std': [0.289, 0.274,...