Some model parameters or buffers are not found in the checkpoint

when i start to train model, the following warning occured: [11/12 11:19:58 d2.checkpoint.detection_checkpoint]: [DetectionCheckpointer] Loading from ./pretrained/dino_vit_s_fna.pth ... [11/12 11:19:58 fvcore.common.checkpoint]: [Checkpointer] Loading from ./pretrained/dino_vit_s_fna.pth ... WARNING [11/12 11:19:58 fvcore.common.checkpoint]: Some model parameters or buffers are not found in the checkpoint: backbone.blocks.0.attn.{rel_pos_h, rel_pos_w} backbone.blocks.1.attn.{rel_pos_h, rel_pos_w} backbone.blocks.10.attn.{rel_pos_h, rel_pos_w} backbone.blocks.11.attn.{rel_pos_h, rel_pos_w} backbone.blocks.11.residual.conv1.weight backbone.blocks.11.residual.conv2.weight backbone.blocks.11.residual.conv3.weight backbone.blocks.11.residual.norm1.{bias, weight} backbone.blocks.11.residual.norm2.{bias, weight} backbone.blocks.11.residual.norm3.{bias, weight} backbone.blocks.2.attn.{rel_pos_h, rel_pos_w} backbone.blocks.2.residual.conv1.weight backbone.blocks.2.residual.conv2.weight backbone.blocks.2.residual.conv3.weight backbone.blocks.2.residual.norm1.{bias, weight} backbone.blocks.2.residual.norm2.{bias, weight} backbone.blocks.2.residual.norm3.{bias, weight} backbone.blocks.3.attn.{rel_pos_h, rel_pos_w} backbone.blocks.4.attn.{rel_pos_h, rel_pos_w} backbone.blocks.5.attn.{rel_pos_h, rel_pos_w} backbone.blocks.5.residual.conv1.weight backbone.blocks.5.residual.conv2.weight backbone.blocks.5.residual.conv3.weight backbone.blocks.5.residual.norm1.{bias, weight} backbone.blocks.5.residual.norm2.{bias, weight} backbone.blocks.5.residual.norm3.{bias, weight} backbone.blocks.6.attn.{rel_pos_h, rel_pos_w} backbone.blocks.7.attn.{rel_pos_h, rel_pos_w} backbone.blocks.8.attn.{rel_pos_h, rel_pos_w} backbone.blocks.8.residual.conv1.weight backbone.blocks.8.residual.conv2.weight backbone.blocks.8.residual.conv3.weight backbone.blocks.8.residual.norm1.{bias, weight} backbone.blocks.8.residual.norm2.{bias, weight} backbone.blocks.8.residual.norm3.{bias, weight} backbone.blocks.9.attn.{rel_pos_h, rel_pos_w} decoder.convstream.convs.0.bn.{bias, running_mean, running_var, weight} decoder.convstream.convs.0.conv.weight decoder.convstream.convs.1.bn.{bias, running_mean, running_var, weight} decoder.convstream.convs.1.conv.weight decoder.convstream.convs.2.bn.{bias, running_mean, running_var, weight} decoder.convstream.convs.2.conv.weight decoder.fusion_blks.0.conv.bn.{bias, running_mean, running_var, weight} decoder.fusion_blks.0.conv.conv.weight decoder.fusion_blks.1.conv.bn.{bias, running_mean, running_var, weight} decoder.fusion_blks.1.conv.conv.weight decoder.fusion_blks.2.conv.bn.{bias, running_mean, running_var, weight} decoder.fusion_blks.2.conv.conv.weight decoder.fusion_blks.3.conv.bn.{bias, running_mean, running_var, weight} decoder.fusion_blks.3.conv.conv.weight decoder.matting_head.matting_convs.0.{bias, weight} decoder.matting_head.matting_convs.1.{bias, running_mean, running_var, weight} decoder.matting_head.matting_convs.3.{bias, weight} WARNING [11/12 11:19:58 fvcore.common.checkpoint]: The checkpoint state_dict contains keys that are not used by the model: backbone.cls_token backbone.norm.{bias, weight} [11/12 11:19:58 d2.engine.train_loop]: Starting training from iteration 0 ep04a_sc000_0960_00002.png does not have enough unknown area for crop. Resized to target size.left_top: (22, 62) ep13a_sc000_2310_00011.png does not have enough unknown area for crop. Resized to target size.left_top: (300, 921) ep02_sc001_0470_00027.png does not have enough unknown area for crop. Resized to target size.left_top: (81, 39) ep02_sc011_0580_00020.png does not have enough unknown area for crop. Resized to target size.left_top: (188, 220) ep06_sc034_0020_00178.png does not have enough unknown area for crop. Resized to target size.left_top: (68, 114)