diff --git a/train/train_poseMF_shapeGaussian_net.py b/train/train_poseMF_shapeGaussian_net.py
index 4b8b445..b6b06bb 100644
--- a/train/train_poseMF_shapeGaussian_net.py
+++ b/train/train_poseMF_shapeGaussian_net.py
@@ -193,18 +193,14 @@ def train_poseMF_shapeGaussian_net(pose_shape_model,
                                                          cam_t=target_cam_t,
                                                          lights_rgb_settings=lights_rgb_settings)
                     
-                    print(f"tran_poseMF_shapeGaussian_net.renderer_output: \n{renderer_output}", flush=True)
-                    
                     iuv_in = renderer_output['iuv_images'].permute(0, 3, 1, 2).contiguous()  # (bs, 3, img_wh, img_wh)
                     iuv_in[:, 1:, :, :] = iuv_in[:, 1:, :, :] * 255
                     iuv_in = iuv_in.round()
                     rgb_in = renderer_output['rgb_images'].permute(0, 3, 1, 2).contiguous()  # (bs, 3, img_wh, img_wh)
 
-                    print(f"train_poseMF_shapeGaussian_net.iuv_in: \n{iuv_in}", flush=True)
                     # Prepare seg for extreme crop augmentation
                     seg_extreme_crop = random_extreme_crop(seg=iuv_in[:, 0, :, :],
                                                            extreme_crop_probability=pose_shape_cfg.TRAIN.SYNTH_DATA.AUGMENT.PROXY_REP.EXTREME_CROP_PROB)
-                    print(f"train_poseMF_shapeGaussian_net.seg_extreme_crop: \n{seg_extreme_crop}", flush=True)
 
                     # Crop to person bounding box after bbox scale and centre augmentation
                     crop_outputs = batch_crop_pytorch_affine(input_wh=(pose_shape_cfg.DATA.PROXY_REP_SIZE, pose_shape_cfg.DATA.PROXY_REP_SIZE),
diff --git a/utils/augmentation/proxy_rep_augmentation.py b/utils/augmentation/proxy_rep_augmentation.py
index 8643aee..7a05619 100644
--- a/utils/augmentation/proxy_rep_augmentation.py
+++ b/utils/augmentation/proxy_rep_augmentation.py
@@ -243,8 +243,6 @@ def random_extreme_crop(seg,
     :param extreme_crop_probability:
     :return: new_seg: part segmentations with regions to extreme crop
     """
-    print(f"proxy_rep_augmentation.random_extreme_crop.seg: \n{seg}", flush=True)
-    
     remove_legs_classes = torch.tensor([5, 6, 7, 8, 9, 10, 11, 12, 13, 14],
                                        device=seg.device,
                                        dtype=seg.dtype)  # Legs and feet