Fixed n_joints bug karfly#73

train.py

@@ -288,10 +288,7 @@ def one_epoch(model, criterion, opt, config, dataloader, device, epoch, n_iters_
                 # batch shape[2] is likely to be the dimension of the keypoints (= 3)
                 # n_views is also the number of cameras being used in this batch 
                 batch_size, n_views, image_shape = images_batch.shape[0], images_batch.shape[1], tuple(images_batch.shape[3:])
-                n_joints = keypoints_3d_pred[0].shape[1]
-                print(n_joints)
-                import ipdb; ipdb.set_trace()
-                print(keypoints_3d_pred[0], keypoints_3d_pred.shape)
+                n_joints = keypoints_3d_pred.shape[1]
 
                 keypoints_3d_binary_validity_gt = (keypoints_3d_validity_gt > 0.0).type(torch.float32)