3D关键点坐标

[ccchen091]

您好，我想请问一下如何通过你这个模型将检测到的3D关键点坐标保存下来呢.

[ChanglongJiangGit]

您好！感谢您的关注！ 如果您需要保存手部的2.5D坐标，那只需要保存'model.py'中的‘pred_keypoints’即可。 由于我们的模型是预测的2.5D坐标，没有预测根节点的绝对坐标或是二者的绝对距离，所以无法得到两手准确的3D坐标。 希望对您有帮助！

[ccchen091]

非常感谢您的回复！ 我还想再问一下您这个2.5D坐标是世界坐标吗？ 同时我还有一个疑惑，就是您这个工作中用的是interhand2.6M数据集，其中的内外参是interhand2.6M数据集给出的吗？

[ChanglongJiangGit]

1、在'model.py'中的‘pred_keypoints’是在图像坐标系下的2.5D坐标，这个参数会反传回dataset.py中进行evaluation，在这个函数中你可以看到跟dataloader的get_item完全相逆的操作，也就是将img坐标系转换到相机坐标系。‘pred_joint_coord_cam’即为相机坐标系下的预测结果，我们也是在这个结果下进行的测试。如果想要世界坐标系下的预测结果，只需要在这个结果的基础上进行cam2world的转换即可。 2、是的，interhand数据集会给出相机的内外参，直接加载即可。

[ccchen091]

感谢您的帮助！祝您工作顺利！

[ccchen091]

您好！又来打扰您了，如果您那有nyu的数据集和hands2017数据集的话可以发给我吗？我在网上找的要不下载不了，要不就是找不到。谢谢！

[ChanglongJiangGit]

您好！ nyu的下载网址：https://jonathantompson.github.io/NYU_Hand_Pose_Dataset.htm#download hands 2017的下载网址：http://icvl.ee.ic.ac.uk/hands17/challenge/ 希望对您有帮助！

[ccchen091]

非常感谢您的回复！！！我自己也有找到这两个网页 要不进不去，要不就是下载不了（我开了梯子的）， 明天我自己再尝试一下！  非常您每次都能很快的回复我.希望您科研顺利!

525935379 @.***

 

------------------ 原始邮件 ------------------ 发件人: "ChanglongJiangGit/A2J-Transformer" @.>; 发送时间: 2024年3月5日(星期二) 晚上9:27 @.>; @.**@.>; 主题: Re: [ChanglongJiangGit/A2J-Transformer] 3D关键点坐标 (Issue #21)

您好！ nyu的下载网址：https://jonathantompson.github.io/NYU_Hand_Pose_Dataset.htm#download hands 2017的下载网址：http://icvl.ee.ic.ac.uk/hands17/challenge/ image.png (view on web) 希望对您有帮助！

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

[ccchen091]

您好，又来打扰您了，我想请问以下我在testNYU数据集的时候加载了A2J论文中给出的NYU.pth权重文件以后出现如下的问题，我搜索后发现应该是权重文件不匹配，所以我想来问一下您加载的model_41_wetD_0.0001_depFact_0.5_RegFact_3_rndShft_5.pth这个权重文件是不是就是NYU.pth，如果不是的话您这个权重文件去哪下载？谢谢！！！ RuntimeError: Error(s) in loading state_dict for DataParallel: Missing key(s) in state_dict: "module.backbone.0.body.conv1.weight", "module.backbone.0.body.bn1.weight", "module.backbone.0.body.bn1.bias", "module.backbone.0.body.bn1.running_mean", "module.backbone.0.body.bn1.running_var", "module.backbone.0.body.layer1.0.conv1.weight", "module.backbone.0.body.layer1.0.bn1.weight", "module.backbone.0.body.layer1.0.bn1.bias", "module.backbone.0.body.layer1.0.bn1.running_mean", "module.backbone.0.body.layer1.0.bn1.running_var", "module.backbone.0.body.layer1.0.conv2.weight", "module.backbone.0.body.layer1.0.bn2.weight", "module.backbone.0.body.layer1.0.bn2.bias", "module.backbone.0.body.layer1.0.bn2.running_mean", "module.backbone.0.body.layer1.0.bn2.running_var", "module.backbone.0.body.layer1.0.conv3.weight", "module.backbone.0.body.layer1.0.bn3.weight", "module.backbone.0.body.layer1.0.bn3.bias", "module.backbone.0.body.layer1.0.bn3.running_mean", "module.backbone.0.body.layer1.0.bn3.running_var", "module.backbone.0.body.layer1.0.downsample.0.weight", "module.backbone.0.body.layer1.0.downsample.1.weight", "module.backbone.0.body.layer1.0.downsample.1.bias", "module.backbone.0.body.layer1.0.downsample.1.running_mean", "module.backbone.0.body.layer1.0.downsample.1.running_var", "module.backbone.0.body.layer1.1.conv1.weight", "module.backbone.0.body.layer1.1.bn1.weight", "module.backbone.0.body.layer1.1.bn1.bias", "module.backbone.0.body.layer1.1.bn1.running_mean", "module.backbone.0.body.layer1.1.bn1.running_var", "module.backbone.0.body.layer1.1.conv2.weight", "module.backbone.0.body.layer1.1.bn2.weight", "module.backbone.0.body.layer1.1.bn2.bias", "module.backbone.0.body.layer1.1.bn2.running_mean", "module.backbone.0.body.layer1.1.bn2.running_var", "module.backbone.0.body.layer1.1.conv3.weight", "module.backbone.0.body.layer1.1.bn3.weight", "module.backbone.0.body.layer1.1.bn3.bias", "module.backbone.0.body.layer1.1.bn3.running_mean", "module.backbone.0.body.layer1.1.bn3.running_var", "module.backbone.0.body.layer1.2.conv1.weight", "module.backbone.0.body.layer1.2.bn1.weight", "module.backbone.0.body.layer1.2.bn1.bias", "module.backbone.0.body.layer1.2.bn1.running_mean", "module.backbone.0.body.layer1.2.bn1.running_var", "module.backbone.0.body.layer1.2.conv2.weight", "module.backbone.0.body.layer1.2.bn2.weight", "module.backbone.0.body.layer1.2.bn2.bias", "module.backbone.0.body.layer1.2.bn2.running_mean", "module.backbone.0.body.layer1.2.bn2.running_var", "module.backbone.0.body.layer1.2.conv3.weight", "module.backbone.0.body.layer1.2.bn3.weight", "module.backbone.0.body.layer1.2.bn3.bias", "module.backbone.0.body.layer1.2.bn3.running_mean", "module.backbone.0.body.layer1.2.bn3.running_var", "module.backbone.0.body.layer2.0.conv1.weight", "module.backbone.0.body.layer2.0.bn1.weight", "module.backbone.0.body.layer2.0.bn1.bias", "module.backbone.0.body.layer2.0.bn1.running_mean", "module.backbone.0.body.layer2.0.bn1.running_var", "module.backbone.0.body.layer2.0.conv2.weight", "module.backbone.0.body.layer2.0.bn2.weight", "module.backbone.0.body.layer2.0.bn2.bias", "module.backbone.0.body.layer2.0.bn2.running_mean", "module.backbone.0.body.layer2.0.bn2.running_var", "module.backbone.0.body.layer2.0.conv3.weight", "module.backbone.0.body.layer2.0.bn3.weight", "module.backbone.0.body.layer2.0.bn3.bias", "module.backbone.0.body.layer2.0.bn3.running_mean", "module.backbone.0.body.layer2.0.bn3.running_var", "module.backbone.0.body.layer2.0.downsample.0.weight", "module.backbone.0.body.layer2.0.downsample.1.weight", "module.backbone.0.body.layer2.0.downsample.1.bias", "module.backbone.0.body.layer2.0.downsample.1.running_mean", "module.backbone.0.body.layer2.0.downsample.1.running_var", "module.backbone.0.body.layer2.1.conv1.weight", "module.backbone.0.body.layer2.1.bn1.weight", "module.backbone.0.body.layer2.1.bn1.bias", "module.backbone.0.body.layer2.1.bn1.running_mean", "module.backbone.0.body.layer2.1.bn1.running_var", "module.backbone.0.body.layer2.1.conv2.weight", "module.backbone.0.body.layer2.1.bn2.weight", "module.backbone.0.body.layer2.1.bn2.bias", "module.backbone.0.body.layer2.1.bn2.running_mean", "module.backbone.0.body.layer2.1.bn2.running_var", "module.backbone.0.body.layer2.1.conv3.weight", "module.backbone.0.body.layer2.1.bn3.weight", "module.backbone.0.body.layer2.1.bn3.bias", "module.backbone.0.body.layer2.1.bn3.running_mean", "module.backbone.0.body.layer2.1.bn3.running_var", "module.backbone.0.body.layer2.2.conv1.weight", "module.backbone.0.body.layer2.2.bn1.weight", "module.backbone.0.body.layer2.2.bn1.bias", "module.backbone.0.body.layer2.2.bn1.running_mean", "module.backbone.0.body.layer2.2.bn1.running_var", "module.backbone.0.body.layer2.2.conv2.weight", "module.backbone.0.body.layer2.2.bn2.weight", "module.backbone.0.body.layer2.2.bn2.bias", "module.backbone.0.body.layer2.2.bn2.running_mean", "module.backbone.0.body.layer2.2.bn2.running_var", "module.backbone.0.body.layer2.2.conv3.weight", "module.backbone.0.body.layer2.2.bn3.weight", "module.backbone.0.body.layer2.2.bn3.bias", "module.backbone.0.body.layer2.2.bn3.running_mean", "module.backbone.0.body.layer2.2.bn3.running_var", "module.backbone.0.body.layer2.3.conv1.weight", "module.backbone.0.body.layer2.3.bn1.weight", "module.backbone.0.body.layer2.3.bn1.bias", "module.backbone.0.body.layer2.3.bn1.running_mean", "module.backbone.0.body.layer2.3.bn1.running_var", "module.backbone.0.body.layer2.3.conv2.weight", "module.backbone.0.body.layer2.3.bn2.weight", "module.backbone.0.body.layer2.3.bn2.bias", "module.backbone.0.body.layer2.3.bn2.running_mean", "module.backbone.0.body.layer2.3.bn2.running_var", "module.backbone.0.body.layer2.3.conv3.weight", "module.backbone.0.body.layer2.3.bn3.weight", "module.backbone.0.body.layer2.3.bn3.bias", "module.backbone.0.body.layer2.3.bn3.running_mean", "module.backbone.0.body.layer2.3.bn3.running_var", "module.backbone.0.body.layer3.0.conv1.weight", "module.backbone.0.body.layer3.0.bn1.weight", "module.backbone.0.body.layer3.0.bn1.bias", "module.backbone.0.body.layer3.0.bn1.running_mean", "module.backbone.0.body.layer3.0.bn1.running_var", "module.backbone.0.body.layer3.0.conv2.weight", "module.backbone.0.body.layer3.0.bn2.weight", "module.backbone.0.body.layer3.0.bn2.bias", "module.backbone.0.body.layer3.0.bn2.running_mean", "module.backbone.0.body.layer3.0.bn2.running_var", "module.backbone.0.body.layer3.0.conv3.weight", "module.backbone.0.body.layer3.0.bn3.weight", "module.backbone.0.body.layer3.0.bn3.bias", "module.backbone.0.body.layer3.0.bn3.running_mean", "module.backbone.0.body.layer3.0.bn3.running_var", "module.backbone.0.body.layer3.0.downsample.0.weight", "module.backbone.0.body.layer3.0.downsample.1.weight", "module.backbone.0.body.layer3.0.downsample.1.bias", "module.backbone.0.body.layer3.0.downsample.1.running_mean", "module.backbone.0.body.layer3.0.downsample.1.running_var", "module.backbone.0.body.layer3.1.conv1.weight", "module.backbone.0.body.layer3.1.bn1.weight", "module.backbone.0.body.layer3.1.bn1.bias", "module.backbone.0.body.layer3.1.bn1.running_mean", "module.backbone.0.body.layer3.1.bn1.running_var", "module.backbone.0.body.layer3.1.conv2.weight", "module.backbone.0.body.layer3.1.bn2.weight", "module.backbone.0.body.layer3.1.bn2.bias", "module.backbone.0.body.layer3.1.bn2.running_mean", "module.backbone.0.body.layer3.1.bn2.running_var", "module.backbone.0.body.layer3.1.conv3.weight", "module.backbone.0.body.layer3.1.bn3.weight", "module.backbone.0.body.layer3.1.bn3.bias", "module.backbone.0.body.layer3.1.bn3.running_mean", "module.backbone.0.body.layer3.1.bn3.running_var", "module.backbone.0.body.layer3.2.conv1.weight", "module.backbone.0.body.layer3.2.bn1.weight", "module.backbone.0.body.layer3.2.bn1.bias", "module.backbone.0.body.layer3.2.bn1.running_mean", "module.backbone.0.body.layer3.2.bn1.running_var", "module.backbone.0.body.layer3.2.conv2.weight", "module.backbone.0.body.layer3.2.bn2.weight", "module.backbone.0.body.layer3.2.bn2.bias", "module.backbone.0.body.layer3.2.bn2.running_mean", "module.backbone.0.body.layer3.2.bn2.running_var", "module.backbone.0.body.layer3.2.conv3.weight", "module.backbone.0.body.layer3.2.bn3.weight", "module.backbone.0.body.layer3.2.bn3.bias", "module.backbone.0.body.layer3.2.bn3.running_mean", "module.backbone.0.body.layer3.2.bn3.running_var", "module.backbone.0.body.layer3.3.conv1.weight", "module.backbone.0.body.layer3.3.bn1.weight", "module.backbone.0.body.layer3.3.bn1.bias", "module.backbone.0.body.layer3.3.bn1.running_mean", "module.backbone.0.body.layer3.3.bn1.running_var", "module.backbone.0.body.layer3.3.conv2.weight", "module.backbone.0.body.layer3.3.bn2.weight", "module.backbone.0.body.layer3.3.bn2.bias", "module.backbone.0.body.layer3.3.bn2.running_mean", "module.backbone.0.body.layer3.3.bn2.running_var", "module.backbone.0.body.layer3.3.conv3.weight", "module.backbone.0.body.layer3.3.bn3.weight", "module.backbone.0.body.layer3.3.bn3.bias", "module.backbone.0.body.layer3.3.bn3.running_mean", "module.backbone.0.body.layer3.3.bn3.running_var", "module.backbone.0.body.layer3.4.conv1.weight", "module.backbone.0.body.layer3.4.bn1.weight", "module.backbone.0.body.layer3.4.bn1.bias", "module.backbone.0.body.layer3.4.bn1.running_mean", "module.backbone.0.body.layer3.4.bn1.running_var", "module.backbone.0.body.layer3.4.conv2.weight", "module.backbone.0.body.layer3.4.bn2.weight", "module.backbone.0.body.layer3.4.bn2.bias", "module.backbone.0.body.layer3.4.bn2.running_mean", "module.backbone.0.body.layer3.4.bn2.running_var", "module.backbone.0.body.layer3.4.conv3.weight", "module.backbone.0.body.layer3.4.bn3.weight", "module.backbone.0.body.layer3.4.bn3.bias", "module.backbone.0.body.layer3.4.bn3.running_mean", "module.backbone.0.body.layer3.4.bn3.running_var", "module.backbone.0.body.layer3.5.conv1.weight", "module.backbone.0.body.layer3.5.bn1.weight", "module.backbone.0.body.layer3.5.bn1.bias", "module.backbone.0.body.layer3.5.bn1.running_mean", "module.backbone.0.body.layer3.5.bn1.running_var", "module.backbone.0.body.layer3.5.conv2.weight", "module.backbone.0.body.layer3.5.bn2.weight", "module.backbone.0.body.layer3.5.bn2.bias", "module.backbone.0.body.layer3.5.bn2.running_mean", "module.backbone.0.body.layer3.5.bn2.running_var", "module.backbone.0.body.layer3.5.conv3.weight", "module.backbone.0.body.layer3.5.bn3.weight", "module.backbone.0.body.layer3.5.bn3.bias", "module.backbone.0.body.layer3.5.bn3.running_mean", "module.backbone.0.body.layer3.5.bn3.running_var", "module.backbone.0.body.layer4.0.conv1.weight", "module.backbone.0.body.layer4.0.bn1.weight", "module.backbone.0.body.layer4.0.bn1.bias", "module.backbone.0.body.layer4.0.bn1.running_mean", "module.backbone.0.body.layer4.0.bn1.running_var", "module.backbone.0.body.layer4.0.conv2.weight", "module.backbone.0.body.layer4.0.bn2.weight", "module.backbone.0.body.layer4.0.bn2.bias", "module.backbone.0.body.layer4.0.bn2.running_mean", "module.backbone.0.body.layer4.0.bn2.running_var", "module.backbone.0.body.layer4.0.conv3.weight", "module.backbone.0.body.layer4.0.bn3.weight", "module.backbone.0.body.layer4.0.bn3.bias", "module.backbone.0.body.layer4.0.bn3.running_mean", "module.backbone.0.body.layer4.0.bn3.running_var", "module.backbone.0.body.layer4.0.downsample.0.weight", "module.backbone.0.body.layer4.0.downsample.1.weight", "module.backbone.0.body.layer4.0.downsample.1.bias", "module.backbone.0.body.layer4.0.downsample.1.running_mean", "module.backbone.0.body.layer4.0.downsample.1.running_var", "module.backbone.0.body.layer4.1.conv1.weight", "module.backbone.0.body.layer4.1.bn1.weight", "module.backbone.0.body.layer4.1.bn1.bias", "module.backbone.0.body.layer4.1.bn1.running_mean", "module.backbone.0.body.layer4.1.bn1.running_var", "module.backbone.0.body.layer4.1.conv2.weight", "module.backbone.0.body.layer4.1.bn2.weight", "module.backbone.0.body.layer4.1.bn2.bias", "module.backbone.0.body.layer4.1.bn2.running_mean", "module.backbone.0.body.layer4.1.bn2.running_var", "module.backbone.0.body.layer4.1.conv3.weight", "module.backbone.0.body.layer4.1.bn3.weight", "module.backbone.0.body.layer4.1.bn3.bias", "module.backbone.0.body.layer4.1.bn3.running_mean", "module.backbone.0.body.layer4.1.bn3.running_var", "module.backbone.0.body.layer4.2.conv1.weight", "module.backbone.0.body.layer4.2.bn1.weight", "module.backbone.0.body.layer4.2.bn1.bias", "module.backbone.0.body.layer4.2.bn1.running_mean", "module.backbone.0.body.layer4.2.bn1.running_var", "module.backbone.0.body.layer4.2.conv2.weight", "module.backbone.0.body.layer4.2.bn2.weight", "module.backbone.0.body.layer4.2.bn2.bias", "module.backbone.0.body.layer4.2.bn2.running_mean", "module.backbone.0.body.layer4.2.bn2.running_var", "module.backbone.0.body.layer4.2.conv3.weight", "module.backbone.0.body.layer4.2.bn3.weight", "module.backbone.0.body.layer4.2.bn3.bias", "module.backbone.0.body.layer4.2.bn3.running_mean", "module.backbone.0.body.layer4.2.bn3.running_var", "module.transformer.level_embed", "module.transformer.encoder.layers.0.self_attn.sampling_offsets.weight", "module.transformer.encoder.layers.0.self_attn.sampling_offsets.bias", "module.transformer.encoder.layers.0.self_attn.attention_weights.weight", "module.transformer.encoder.layers.0.self_attn.attention_weights.bias", "module.transformer.encoder.layers.0.self_attn.value_proj.weight", "module.transformer.encoder.layers.0.self_attn.value_proj.bias", "module.transformer.encoder.layers.0.self_attn.output_proj.weight", "module.transformer.encoder.layers.0.self_attn.output_proj.bias", "module.transformer.encoder.layers.0.norm1.weight", "module.transformer.encoder.layers.0.norm1.bias", "module.transformer.encoder.layers.0.linear1.weight", "module.transformer.encoder.layers.0.linear1.bias", "module.transformer.encoder.layers.0.linear2.weight", "module.transformer.encoder.layers.0.linear2.bias", "module.transformer.encoder.layers.0.norm2.weight", "module.transformer.encoder.layers.0.norm2.bias", "module.transformer.encoder.layers.1.self_attn.sampling_offsets.weight", "module.transformer.encoder.layers.1.self_attn.sampling_offsets.bias", "module.transformer.encoder.layers.1.self_attn.attention_weights.weight", "module.transformer.encoder.layers.1.self_attn.attention_weights.bias", "module.transformer.encoder.layers.1.self_attn.value_proj.weight", "module.transformer.encoder.layers.1.self_attn.value_proj.bias", "module.transformer.encoder.layers.1.self_attn.output_proj.weight", "module.transformer.encoder.layers.1.self_attn.output_proj.bias", "module.transformer.encoder.layers.1.norm1.weight", "module.transformer.encoder.layers.1.norm1.bias", "module.transformer.encoder.layers.1.linear1.weight", "module.transformer.encoder.layers.1.linear1.bias", "module.transformer.encoder.layers.1.linear2.weight", "module.transformer.encoder.layers.1.linear2.bias", "module.transformer.encoder.layers.1.norm2.weight", "module.transformer.encoder.layers.1.norm2.bias", "module.transformer.encoder.layers.2.self_attn.sampling_offsets.weight", "module.transformer.encoder.layers.2.self_attn.sampling_offsets.bias", "module.transformer.encoder.layers.2.self_attn.attention_weights.weight", "module.transformer.encoder.layers.2.self_attn.attention_weights.bias", "module.transformer.encoder.layers.2.self_attn.value_proj.weight", "module.transformer.encoder.layers.2.self_attn.value_proj.bias", "module.transformer.encoder.layers.2.self_attn.output_proj.weight", "module.transformer.encoder.layers.2.self_attn.output_proj.bias", "module.transformer.encoder.layers.2.norm1.weight", "module.transformer.encoder.layers.2.norm1.bias", "module.transformer.encoder.layers.2.linear1.weight", "module.transformer.encoder.layers.2.linear1.bias", "module.transformer.encoder.layers.2.linear2.weight", "module.transformer.encoder.layers.2.linear2.bias", "module.transformer.encoder.layers.2.norm2.weight", "module.transformer.encoder.layers.2.norm2.bias", "module.transformer.encoder.layers.3.self_attn.sampling_offsets.weight", "module.transformer.encoder.layers.3.self_attn.sampling_offsets.bias", "module.transformer.encoder.layers.3.self_attn.attention_weights.weight", "module.transformer.encoder.layers.3.self_attn.attention_weights.bias", "module.transformer.encoder.layers.3.self_attn.value_proj.weight", "module.transformer.encoder.layers.3.self_attn.value_proj.bias", "module.transformer.encoder.layers.3.self_attn.output_proj.weight", "module.transformer.encoder.layers.3.self_attn.output_proj.bias", "module.transformer.encoder.layers.3.norm1.weight", "module.transformer.encoder.layers.3.norm1.bias", "module.transformer.encoder.layers.3.linear1.weight", "module.transformer.encoder.layers.3.linear1.bias", "module.transformer.encoder.layers.3.linear2.weight", "module.transformer.encoder.layers.3.linear2.bias", "module.transformer.encoder.layers.3.norm2.weight", "module.transformer.encoder.layers.3.norm2.bias", "module.transformer.encoder.layers.4.self_attn.sampling_offsets.weight", "module.transformer.encoder.layers.4.self_attn.sampling_offsets.bias", "module.transformer.encoder.layers.4.self_attn.attention_weights.weight", "module.transformer.encoder.layers.4.self_attn.attention_weights.bias", "module.transformer.encoder.layers.4.self_attn.value_proj.weight", "module.transformer.encoder.layers.4.self_attn.value_proj.bias", "module.transformer.encoder.layers.4.self_attn.output_proj.weight", "module.transformer.encoder.layers.4.self_attn.output_proj.bias", "module.transformer.encoder.layers.4.norm1.weight", "module.transformer.encoder.layers.4.norm1.bias", "module.transformer.encoder.layers.4.linear1.weight", "module.transformer.encoder.layers.4.linear1.bias", "module.transformer.encoder.layers.4.linear2.weight", "module.transformer.encoder.layers.4.linear2.bias", "module.transformer.encoder.layers.4.norm2.weight", "module.transformer.encoder.layers.4.norm2.bias", "module.transformer.encoder.layers.5.self_attn.sampling_offsets.weight", "module.transformer.encoder.layers.5.self_attn.sampling_offsets.bias", "module.transformer.encoder.layers.5.self_attn.attention_weights.weight", "module.transformer.encoder.layers.5.self_attn.attention_weights.bias", "module.transformer.encoder.layers.5.self_attn.value_proj.weight", "module.transformer.encoder.layers.5.self_attn.value_proj.bias", "module.transformer.encoder.layers.5.self_attn.output_proj.weight", "module.transformer.encoder.layers.5.self_attn.output_proj.bias", "module.transformer.encoder.layers.5.norm1.weight", "module.transformer.encoder.layers.5.norm1.bias", "module.transformer.encoder.layers.5.linear1.weight", "module.transformer.encoder.layers.5.linear1.bias", "module.transformer.encoder.layers.5.linear2.weight", "module.transformer.encoder.layers.5.linear2.bias", "module.transformer.encoder.layers.5.norm2.weight", "module.transformer.encoder.layers.5.norm2.bias", "module.transformer.decoder.layers.0.cross_attn.sampling_offsets.weight", "module.transformer.decoder.layers.0.cross_attn.sampling_offsets.bias", "module.transformer.decoder.layers.0.cross_attn.attention_weights.weight", "module.transformer.decoder.layers.0.cross_attn.attention_weights.bias", "module.transformer.decoder.layers.0.cross_attn.value_proj.weight", "module.transformer.decoder.layers.0.cross_attn.value_proj.bias", "module.transformer.decoder.layers.0.cross_attn.output_proj.weight", "module.transformer.decoder.layers.0.cross_attn.output_proj.bias", "module.transformer.decoder.layers.0.norm1.weight", "module.transformer.decoder.layers.0.norm1.bias", "module.transformer.decoder.layers.0.self_attn.in_proj_weight", "module.transformer.decoder.layers.0.self_attn.in_proj_bias", "module.transformer.decoder.layers.0.self_attn.out_proj.weight", "module.transformer.decoder.layers.0.self_attn.out_proj.bias", "module.transformer.decoder.layers.0.norm2.weight", "module.transformer.decoder.layers.0.norm2.bias", "module.transformer.decoder.layers.0.linear1.weight", "module.transformer.decoder.layers.0.linear1.bias", "module.transformer.decoder.layers.0.linear2.weight", "module.transformer.decoder.layers.0.linear2.bias", "module.transformer.decoder.layers.0.norm3.weight", "module.transformer.decoder.layers.0.norm3.bias", "module.transformer.decoder.layers.1.cross_attn.sampling_offsets.weight", "module.transformer.decoder.layers.1.cross_attn.sampling_offsets.bias", "module.transformer.decoder.layers.1.cross_attn.attention_weights.weight", "module.transformer.decoder.layers.1.cross_attn.attention_weights.bias", "module.transformer.decoder.layers.1.cross_attn.value_proj.weight", "module.transformer.decoder.layers.1.cross_attn.value_proj.bias", "module.transformer.decoder.layers.1.cross_attn.output_proj.weight", "module.transformer.decoder.layers.1.cross_attn.output_proj.bias", "module.transformer.decoder.layers.1.norm1.weight", "module.transformer.decoder.layers.1.norm1.bias", "module.transformer.decoder.layers.1.self_attn.in_proj_weight", "module.transformer.decoder.layers.1.self_attn.in_proj_bias", "module.transformer.decoder.layers.1.self_attn.out_proj.weight", "module.transformer.decoder.layers.1.self_attn.out_proj.bias", "module.transformer.decoder.layers.1.norm2.weight", "module.transformer.decoder.layers.1.norm2.bias", "module.transformer.decoder.layers.1.linear1.weight", "module.transformer.decoder.layers.1.linear1.bias", "module.transformer.decoder.layers.1.linear2.weight", "module.transformer.decoder.layers.1.linear2.bias", "module.transformer.decoder.layers.1.norm3.weight", "module.transformer.decoder.layers.1.norm3.bias", "module.transformer.decoder.layers.2.cross_attn.sampling_offsets.weight", "module.transformer.decoder.layers.2.cross_attn.sampling_offsets.bias", "module.transformer.decoder.layers.2.cross_attn.attention_weights.weight", "module.transformer.decoder.layers.2.cross_attn.attention_weights.bias", "module.transformer.decoder.layers.2.cross_attn.value_proj.weight", "module.transformer.decoder.layers.2.cross_attn.value_proj.bias", "module.transformer.decoder.layers.2.cross_attn.output_proj.weight", "module.transformer.decoder.layers.2.cross_attn.output_proj.bias", "module.transformer.decoder.layers.2.norm1.weight", "module.transformer.decoder.layers.2.norm1.bias", "module.transformer.decoder.layers.2.self_attn.in_proj_weight", "module.transformer.decoder.layers.2.self_attn.in_proj_bias", "module.transformer.decoder.layers.2.self_attn.out_proj.weight", "module.transformer.decoder.layers.2.self_attn.out_proj.bias", "module.transformer.decoder.layers.2.norm2.weight", "module.transformer.decoder.layers.2.norm2.bias", "module.transformer.decoder.layers.2.linear1.weight", "module.transformer.decoder.layers.2.linear1.bias", "module.transformer.decoder.layers.2.linear2.weight", "module.transformer.decoder.layers.2.linear2.bias", "module.transformer.decoder.layers.2.norm3.weight", "module.transformer.decoder.layers.2.norm3.bias", "module.transformer.decoder.layers.3.cross_attn.sampling_offsets.weight", "module.transformer.decoder.layers.3.cross_attn.sampling_offsets.bias", "module.transformer.decoder.layers.3.cross_attn.attention_weights.weight", "module.transformer.decoder.layers.3.cross_attn.attention_weights.bias", "module.transformer.decoder.layers.3.cross_attn.value_proj.weight", "module.transformer.decoder.layers.3.cross_attn.value_proj.bias", "module.transformer.decoder.layers.3.cross_attn.output_proj.weight", "module.transformer.decoder.layers.3.cross_attn.output_proj.bias", "module.transformer.decoder.layers.3.norm1.weight", "module.transformer.decoder.layers.3.norm1.bias", "module.transformer.decoder.layers.3.self_attn.in_proj_weight", "module.transformer.decoder.layers.3.self_attn.in_proj_bias", "module.transformer.decoder.layers.3.self_attn.out_proj.weight", "module.transformer.decoder.layers.3.self_attn.out_proj.bias", "module.transformer.decoder.layers.3.norm2.weight", "module.transformer.decoder.layers.3.norm2.bias", "module.transformer.decoder.layers.3.linear1.weight", "module.transformer.decoder.layers.3.linear1.bias", "module.transformer.decoder.layers.3.linear2.weight", "module.transformer.decoder.layers.3.linear2.bias", "module.transformer.decoder.layers.3.norm3.weight", "module.transformer.decoder.layers.3.norm3.bias", "module.transformer.decoder.layers.4.cross_attn.sampling_offsets.weight", "module.transformer.decoder.layers.4.cross_attn.sampling_offsets.bias", "module.transformer.decoder.layers.4.cross_attn.attention_weights.weight", "module.transformer.decoder.layers.4.cross_attn.attention_weights.bias", "module.transformer.decoder.layers.4.cross_attn.value_proj.weight", "module.transformer.decoder.layers.4.cross_attn.value_proj.bias", "module.transformer.decoder.layers.4.cross_attn.output_proj.weight", "module.transformer.decoder.layers.4.cross_attn.output_proj.bias", "module.transformer.decoder.layers.4.norm1.weight", "module.transformer.decoder.layers.4.norm1.bias", "module.transformer.decoder.layers.4.self_attn.in_proj_weight", "module.transformer.decoder.layers.4.self_attn.in_proj_bias", "module.transformer.decoder.layers.4.self_attn.out_proj.weight", "module.transformer.decoder.layers.4.self_attn.out_proj.bias", "module.transformer.decoder.layers.4.norm2.weight", "module.transformer.decoder.layers.4.norm2.bias", "module.transformer.decoder.layers.4.linear1.weight", "module.transformer.decoder.layers.4.linear1.bias", "module.transformer.decoder.layers.4.linear2.weight", "module.transformer.decoder.layers.4.linear2.bias", "module.transformer.decoder.layers.4.norm3.weight", "module.transformer.decoder.layers.4.norm3.bias", "module.transformer.decoder.layers.5.cross_attn.sampling_offsets.weight", "module.transformer.decoder.layers.5.cross_attn.sampling_offsets.bias", "module.transformer.decoder.layers.5.cross_attn.attention_weights.weight", "module.transformer.decoder.layers.5.cross_attn.attention_weights.bias", "module.transformer.decoder.layers.5.cross_attn.value_proj.weight", "module.transformer.decoder.layers.5.cross_attn.value_proj.bias", "module.transformer.decoder.layers.5.cross_attn.output_proj.weight", "module.transformer.decoder.layers.5.cross_attn.output_proj.bias", "module.transformer.decoder.layers.5.norm1.weight", "module.transformer.decoder.layers.5.norm1.bias", "module.transformer.decoder.layers.5.self_attn.in_proj_weight", "module.transformer.decoder.layers.5.self_attn.in_proj_bias", "module.transformer.decoder.layers.5.self_attn.out_proj.weight", "module.transformer.decoder.layers.5.self_attn.out_proj.bias", "module.transformer.decoder.layers.5.norm2.weight", "module.transformer.decoder.layers.5.norm2.bias", "module.transformer.decoder.layers.5.linear1.weight", "module.transformer.decoder.layers.5.linear1.bias", "module.transformer.decoder.layers.5.linear2.weight", "module.transformer.decoder.layers.5.linear2.bias", "module.transformer.decoder.layers.5.norm3.weight", "module.transformer.decoder.layers.5.norm3.bias", "module.transformer.decoder.query_scale.layers.0.weight", "module.transformer.decoder.query_scale.layers.0.bias", "module.transformer.decoder.query_scale.layers.1.weight", "module.transformer.decoder.query_scale.layers.1.bias", "module.transformer.decoder.ref_point_head.layers.0.weight", "module.transformer.decoder.ref_point_head.layers.0.bias", "module.transformer.decoder.ref_point_head.layers.1.weight", "module.transformer.decoder.ref_point_head.layers.1.bias", "module.transformer.decoder.bbox_embed.0.layers.0.weight", "module.transformer.decoder.bbox_embed.0.layers.0.bias", "module.transformer.decoder.bbox_embed.0.layers.1.weight", "module.transformer.decoder.bbox_embed.0.layers.1.bias", "module.transformer.decoder.bbox_embed.0.layers.2.weight", "module.transformer.decoder.bbox_embed.0.layers.2.bias", "module.transformer.decoder.bbox_embed.1.layers.0.weight", "module.transformer.decoder.bbox_embed.1.layers.0.bias", "module.transformer.decoder.bbox_embed.1.layers.1.weight", "module.transformer.decoder.bbox_embed.1.layers.1.bias", "module.transformer.decoder.bbox_embed.1.layers.2.weight", "module.transformer.decoder.bbox_embed.1.layers.2.bias", "module.transformer.decoder.bbox_embed.2.layers.0.weight", "module.transformer.decoder.bbox_embed.2.layers.0.bias", "module.transformer.decoder.bbox_embed.2.layers.1.weight", "module.transformer.decoder.bbox_embed.2.layers.1.bias", "module.transformer.decoder.bbox_embed.2.layers.2.weight", "module.transformer.decoder.bbox_embed.2.layers.2.bias", "module.transformer.decoder.bbox_embed.3.layers.0.weight", "module.transformer.decoder.bbox_embed.3.layers.0.bias", "module.transformer.decoder.bbox_embed.3.layers.1.weight", "module.transformer.decoder.bbox_embed.3.layers.1.bias", "module.transformer.decoder.bbox_embed.3.layers.2.weight", "module.transformer.decoder.bbox_embed.3.layers.2.bias", "module.transformer.decoder.bbox_embed.4.layers.0.weight", "module.transformer.decoder.bbox_embed.4.layers.0.bias", "module.transformer.decoder.bbox_embed.4.layers.1.weight", "module.transformer.decoder.bbox_embed.4.layers.1.bias", "module.transformer.decoder.bbox_embed.4.layers.2.weight", "module.transformer.decoder.bbox_embed.4.layers.2.bias", "module.transformer.decoder.bbox_embed.5.layers.0.weight", "module.transformer.decoder.bbox_embed.5.layers.0.bias", "module.transformer.decoder.bbox_embed.5.layers.1.weight", "module.transformer.decoder.bbox_embed.5.layers.1.bias", "module.transformer.decoder.bbox_embed.5.layers.2.weight", "module.transformer.decoder.bbox_embed.5.layers.2.bias", "module.bbox_embed_anchor.0.layers.0.weight", "module.bbox_embed_anchor.0.layers.0.bias", "module.bbox_embed_anchor.0.layers.1.weight", "module.bbox_embed_anchor.0.layers.1.bias", "module.bbox_embed_anchor.0.layers.2.weight", "module.bbox_embed_anchor.0.layers.2.bias", "module.bbox_embed_anchor.1.layers.0.weight", "module.bbox_embed_anchor.1.layers.0.bias", "module.bbox_embed_anchor.1.layers.1.weight", "module.bbox_embed_anchor.1.layers.1.bias", "module.bbox_embed_anchor.1.layers.2.weight", "module.bbox_embed_anchor.1.layers.2.bias", "module.bbox_embed_anchor.2.layers.0.weight", "module.bbox_embed_anchor.2.layers.0.bias", "module.bbox_embed_anchor.2.layers.1.weight", "module.bbox_embed_anchor.2.layers.1.bias", "module.bbox_embed_anchor.2.layers.2.weight", "module.bbox_embed_anchor.2.layers.2.bias", "module.bbox_embed_anchor.3.layers.0.weight", "module.bbox_embed_anchor.3.layers.0.bias", "module.bbox_embed_anchor.3.layers.1.weight", "module.bbox_embed_anchor.3.layers.1.bias", "module.bbox_embed_anchor.3.layers.2.weight", "module.bbox_embed_anchor.3.layers.2.bias", "module.bbox_embed_anchor.4.layers.0.weight", "module.bbox_embed_anchor.4.layers.0.bias", "module.bbox_embed_anchor.4.layers.1.weight", "module.bbox_embed_anchor.4.layers.1.bias", "module.bbox_embed_anchor.4.layers.2.weight", "module.bbox_embed_anchor.4.layers.2.bias", "module.bbox_embed_anchor.5.layers.0.weight", "module.bbox_embed_anchor.5.layers.0.bias", "module.bbox_embed_anchor.5.layers.1.weight", "module.bbox_embed_anchor.5.layers.1.bias", "module.bbox_embed_anchor.5.layers.2.weight", "module.bbox_embed_anchor.5.layers.2.bias", "module.bbox_embed_keypoints.0.layers.0.weight", "module.bbox_embed_keypoints.0.layers.0.bias", "module.bbox_embed_keypoints.0.layers.1.weight", "module.bbox_embed_keypoints.0.layers.1.bias", "module.bbox_embed_keypoints.0.layers.2.weight", "module.bbox_embed_keypoints.0.layers.2.bias", "module.bbox_embed_keypoints.1.layers.0.weight", "module.bbox_embed_keypoints.1.layers.0.bias", "module.bbox_embed_keypoints.1.layers.1.weight", "module.bbox_embed_keypoints.1.layers.1.bias", "module.bbox_embed_keypoints.1.layers.2.weight", "module.bbox_embed_keypoints.1.layers.2.bias", "module.bbox_embed_keypoints.2.layers.0.weight", "module.bbox_embed_keypoints.2.layers.0.bias", "module.bbox_embed_keypoints.2.layers.1.weight", "module.bbox_embed_keypoints.2.layers.1.bias", "module.bbox_embed_keypoints.2.layers.2.weight", "module.bbox_embed_keypoints.2.layers.2.bias", "module.bbox_embed_keypoints.3.layers.0.weight", "module.bbox_embed_keypoints.3.layers.0.bias", "module.bbox_embed_keypoints.3.layers.1.weight", "module.bbox_embed_keypoints.3.layers.1.bias", "module.bbox_embed_keypoints.3.layers.2.weight", "module.bbox_embed_keypoints.3.layers.2.bias", "module.bbox_embed_keypoints.4.layers.0.weight", "module.bbox_embed_keypoints.4.layers.0.bias", "module.bbox_embed_keypoints.4.layers.1.weight", "module.bbox_embed_keypoints.4.layers.1.bias", "module.bbox_embed_keypoints.4.layers.2.weight", "module.bbox_embed_keypoints.4.layers.2.bias", "module.bbox_embed_keypoints.5.layers.0.weight", "module.bbox_embed_keypoints.5.layers.0.bias", "module.bbox_embed_keypoints.5.layers.1.weight", "module.bbox_embed_keypoints.5.layers.1.bias", "module.bbox_embed_keypoints.5.layers.2.weight", "module.bbox_embed_keypoints.5.layers.2.bias", "module.anchor_weights.0.layers.0.weight", "module.anchor_weights.0.layers.0.bias", "module.anchor_weights.0.layers.1.weight", "module.anchor_weights.0.layers.1.bias", "module.anchor_weights.0.layers.2.weight", "module.anchor_weights.0.layers.2.bias", "module.anchor_weights.1.layers.0.weight", "module.anchor_weights.1.layers.0.bias", "module.anchor_weights.1.layers.1.weight", "module.anchor_weights.1.layers.1.bias", "module.anchor_weights.1.layers.2.weight", "module.anchor_weights.1.layers.2.bias", "module.anchor_weights.2.layers.0.weight", "module.anchor_weights.2.layers.0.bias", "module.anchor_weights.2.layers.1.weight", "module.anchor_weights.2.layers.1.bias", "module.anchor_weights.2.layers.2.weight", "module.anchor_weights.2.layers.2.bias", "module.anchor_weights.3.layers.0.weight", "module.anchor_weights.3.layers.0.bias", "module.anchor_weights.3.layers.1.weight", "module.anchor_weights.3.layers.1.bias", "module.anchor_weights.3.layers.2.weight", "module.anchor_weights.3.layers.2.bias", "module.anchor_weights.4.layers.0.weight", "module.anchor_weights.4.layers.0.bias", "module.anchor_weights.4.layers.1.weight", "module.anchor_weights.4.layers.1.bias", "module.anchor_weights.4.layers.2.weight", "module.anchor_weights.4.layers.2.bias", "module.anchor_weights.5.layers.0.weight", "module.anchor_weights.5.layers.0.bias", "module.anchor_weights.5.layers.1.weight", "module.anchor_weights.5.layers.1.bias", "module.anchor_weights.5.layers.2.weight", "module.anchor_weights.5.layers.2.bias", "module.tgt_embed.weight", "module.refpoint_embed.weight", "module.input_proj.0.0.weight", "module.input_proj.0.0.bias", "module.input_proj.0.1.weight", "module.input_proj.0.1.bias", "module.input_proj.1.0.weight", "module.input_proj.1.0.bias", "module.input_proj.1.1.weight", "module.input_proj.1.1.bias", "module.input_proj.2.0.weight", "module.input_proj.2.0.bias", "module.input_proj.2.1.weight", "module.input_proj.2.1.bias", "module.input_proj.3.0.weight", "module.input_proj.3.0.bias", "module.input_proj.3.1.weight", "module.input_proj.3.1.bias". Unexpected key(s) in state_dict: "Backbone.model.conv1.weight", "Backbone.model.bn1.weight", "Backbone.model.bn1.bias", "Backbone.model.bn1.running_mean", "Backbone.model.bn1.running_var", "Backbone.model.bn1.num_batches_tracked", "Backbone.model.layer1.0.conv1.weight", "Backbone.model.layer1.0.bn1.weight", "Backbone.model.layer1.0.bn1.bias", "Backbone.model.layer1.0.bn1.running_mean", "Backbone.model.layer1.0.bn1.running_var", "Backbone.model.layer1.0.bn1.num_batches_tracked", "Backbone.model.layer1.0.conv2.weight", "Backbone.model.layer1.0.bn2.weight", "Backbone.model.layer1.0.bn2.bias", "Backbone.model.layer1.0.bn2.running_mean", "Backbone.model.layer1.0.bn2.running_var", "Backbone.model.layer1.0.bn2.num_batches_tracked", "Backbone.model.layer1.0.conv3.weight", "Backbone.model.layer1.0.bn3.weight", "Backbone.model.layer1.0.bn3.bias", "Backbone.model.layer1.0.bn3.running_mean", "Backbone.model.layer1.0.bn3.running_var", "Backbone.model.layer1.0.bn3.num_batches_tracked", "Backbone.model.layer1.0.downsample.0.weight", "Backbone.model.layer1.0.downsample.1.weight", "Backbone.model.layer1.0.downsample.1.bias", "Backbone.model.layer1.0.downsample.1.running_mean", "Backbone.model.layer1.0.downsample.1.running_var", "Backbone.model.layer1.0.downsample.1.num_batches_tracked", "Backbone.model.layer1.1.conv1.weight", "Backbone.model.layer1.1.bn1.weight", "Backbone.model.layer1.1.bn1.bias", "Backbone.model.layer1.1.bn1.running_mean", "Backbone.model.layer1.1.bn1.running_var", "Backbone.model.layer1.1.bn1.num_batches_tracked", "Backbone.model.layer1.1.conv2.weight", "Backbone.model.layer1.1.bn2.weight", "Backbone.model.layer1.1.bn2.bias", "Backbone.model.layer1.1.bn2.running_mean", "Backbone.model.layer1.1.bn2.running_var", "Backbone.model.layer1.1.bn2.num_batches_tracked", "Backbone.model.layer1.1.conv3.weight", "Backbone.model.layer1.1.bn3.weight", "Backbone.model.layer1.1.bn3.bias", "Backbone.model.layer1.1.bn3.running_mean", "Backbone.model.layer1.1.bn3.running_var", "Backbone.model.layer1.1.bn3.num_batches_tracked", "Backbone.model.layer1.2.conv1.weight", "Backbone.model.layer1.2.bn1.weight", "Backbone.model.layer1.2.bn1.bias", "Backbone.model.layer1.2.bn1.running_mean", "Backbone.model.layer1.2.bn1.running_var", "Backbone.model.layer1.2.bn1.num_batches_tracked", "Backbone.model.layer1.2.conv2.weight", "Backbone.model.layer1.2.bn2.weight", "Backbone.model.layer1.2.bn2.bias", "Backbone.model.layer1.2.bn2.running_mean", "Backbone.model.layer1.2.bn2.running_var", "Backbone.model.layer1.2.bn2.num_batches_tracked", "Backbone.model.layer1.2.conv3.weight", "Backbone.model.layer1.2.bn3.weight", "Backbone.model.layer1.2.bn3.bias", "Backbone.model.layer1.2.bn3.running_mean", "Backbone.model.layer1.2.bn3.running_var", "Backbone.model.layer1.2.bn3.num_batches_tracked", "Backbone.model.layer2.0.conv1.weight", "Backbone.model.layer2.0.bn1.weight", "Backbone.model.layer2.0.bn1.bias", "Backbone.model.layer2.0.bn1.running_mean", "Backbone.model.layer2.0.bn1.running_var", "Backbone.model.layer2.0.bn1.num_batches_tracked", "Backbone.model.layer2.0.conv2.weight", "Backbone.model.layer2.0.bn2.weight", "Backbone.model.layer2.0.bn2.bias", "Backbone.model.layer2.0.bn2.running_mean", "Backbone.model.layer2.0.bn2.running_var", "Backbone.model.layer2.0.bn2.num_batches_tracked", "Backbone.model.layer2.0.conv3.weight", "Backbone.model.layer2.0.bn3.weight", "Backbone.model.layer2.0.bn3.bias", "Backbone.model.layer2.0.bn3.running_mean", "Backbone.model.layer2.0.bn3.running_var", "Backbone.model.layer2.0.bn3.num_batches_tracked", "Backbone.model.layer2.0.downsample.0.weight", "Backbone.model.layer2.0.downsample.1.weight", "Backbone.model.layer2.0.downsample.1.bias", "Backbone.model.layer2.0.downsample.1.running_mean", "Backbone.model.layer2.0.downsample.1.running_var", "Backbone.model.layer2.0.downsample.1.num_batches_tracked", "Backbone.model.layer2.1.conv1.weight", "Backbone.model.layer2.1.bn1.weight", "Backbone.model.layer2.1.bn1.bias", "Backbone.model.layer2.1.bn1.running_mean", "Backbone.model.layer2.1.bn1.running_var", "Backbone.model.layer2.1.bn1.num_batches_tracked", "Backbone.model.layer2.1.conv2.weight", "Backbone.model.layer2.1.bn2.weight", "Backbone.model.layer2.1.bn2.bias", "Backbone.model.layer2.1.bn2.running_mean", "Backbone.model.layer2.1.bn2.running_var", "Backbone.model.layer2.1.bn2.num_batches_tracked", "Backbone.model.layer2.1.conv3.weight", "Backbone.model.layer2.1.bn3.weight", "Backbone.model.layer2.1.bn3.bias", "Backbone.model.layer2.1.bn3.running_mean", "Backbone.model.layer2.1.bn3.running_var", "Backbone.model.layer2.1.bn3.num_batches_tracked", "Backbone.model.layer2.2.conv1.weight", "Backbone.model.layer2.2.bn1.weight", "Backbone.model.layer2.2.bn1.bias", "Backbone.model.layer2.2.bn1.running_mean", "Backbone.model.layer2.2.bn1.running_var", "Backbone.model.layer2.2.bn1.num_batches_tracked", "Backbone.model.layer2.2.conv2.weight", "Backbone.model.layer2.2.bn2.weight", "Backbone.model.layer2.2.bn2.bias", "Backbone.model.layer2.2.bn2.running_mean", "Backbone.model.layer2.2.bn2.running_var", "Backbone.model.layer2.2.bn2.num_batches_tracked", "Backbone.model.layer2.2.conv3.weight", "Backbone.model.layer2.2.bn3.weight", "Backbone.model.layer2.2.bn3.bias", "Backbone.model.layer2.2.bn3.running_mean", "Backbone.model.layer2.2.bn3.running_var", "Backbone.model.layer2.2.bn3.num_batches_tracked", "Backbone.model.layer2.3.conv1.weight", "Backbone.model.layer2.3.bn1.weight", "Backbone.model.layer2.3.bn1.bias", "Backbone.model.layer2.3.bn1.running_mean", "Backbone.model.layer2.3.bn1.running_var", "Backbone.model.layer2.3.bn1.num_batches_tracked", "Backbone.model.layer2.3.conv2.weight", "Backbone.model.layer2.3.bn2.weight", "Backbone.model.layer2.3.bn2.bias", "Backbone.model.layer2.3.bn2.running_mean", "Backbone.model.layer2.3.bn2.running_var", "Backbone.model.layer2.3.bn2.num_batches_tracked", "Backbone.model.layer2.3.conv3.weight", "Backbone.model.layer2.3.bn3.weight", "Backbone.model.layer2.3.bn3.bias", "Backbone.model.layer2.3.bn3.running_mean", "Backbone.model.layer2.3.bn3.running_var", "Backbone.model.layer2.3.bn3.num_batches_tracked", "Backbone.model.layer3.0.conv1.weight", "Backbone.model.layer3.0.bn1.weight", "Backbone.model.layer3.0.bn1.bias", "Backbone.model.layer3.0.bn1.running_mean", "Backbone.model.layer3.0.bn1.running_var", "Backbone.model.layer3.0.bn1.num_batches_tracked", "Backbone.model.layer3.0.conv2.weight", "Backbone.model.layer3.0.bn2.weight", "Backbone.model.layer3.0.bn2.bias", "Backbone.model.layer3.0.bn2.running_mean", "Backbone.model.layer3.0.bn2.running_var", "Backbone.model.layer3.0.bn2.num_batches_tracked", "Backbone.model.layer3.0.conv3.weight", "Backbone.model.layer3.0.bn3.weight", "Backbone.model.layer3.0.bn3.bias", "Backbone.model.layer3.0.bn3.running_mean", "Backbone.model.layer3.0.bn3.running_var", "Backbone.model.layer3.0.bn3.num_batches_tracked", "Backbone.model.layer3.0.downsample.0.weight", "Backbone.model.layer3.0.downsample.1.weight", "Backbone.model.layer3.0.downsample.1.bias", "Backbone.model.layer3.0.downsample.1.running_mean", "Backbone.model.layer3.0.downsample.1.running_var", "Backbone.model.layer3.0.downsample.1.num_batches_tracked", "Backbone.model.layer3.1.conv1.weight", "Backbone.model.layer3.1.bn1.weight", "Backbone.model.layer3.1.bn1.bias", "Backbone.model.layer3.1.bn1.running_mean", "Backbone.model.layer3.1.bn1.running_var", "Backbone.model.layer3.1.bn1.num_batches_tracked", "Backbone.model.layer3.1.conv2.weight", "Backbone.model.layer3.1.bn2.weight", "Backbone.model.layer3.1.bn2.bias", "Backbone.model.layer3.1.bn2.running_mean", "Backbone.model.layer3.1.bn2.running_var", "Backbone.model.layer3.1.bn2.num_batches_tracked", "Backbone.model.layer3.1.conv3.weight", "Backbone.model.layer3.1.bn3.weight", "Backbone.model.layer3.1.bn3.bias", "Backbone.model.layer3.1.bn3.running_mean", "Backbone.model.layer3.1.bn3.running_var", "Backbone.model.layer3.1.bn3.num_batches_tracked", "Backbone.model.layer3.2.conv1.weight", "Backbone.model.layer3.2.bn1.weight", "Backbone.model.layer3.2.bn1.bias", "Backbone.model.layer3.2.bn1.running_mean", "Backbone.model.layer3.2.bn1.running_var", "Backbone.model.layer3.2.bn1.num_batches_tracked", "Backbone.model.layer3.2.conv2.weight", "Backbone.model.layer3.2.bn2.weight", "Backbone.model.layer3.2.bn2.bias", "Backbone.model.layer3.2.bn2.running_mean", "Backbone.model.layer3.2.bn2.running_var", "Backbone.model.layer3.2.bn2.num_batches_tracked", "Backbone.model.layer3.2.conv3.weight", "Backbone.model.layer3.2.bn3.weight", "Backbone.model.layer3.2.bn3.bias", "Backbone.model.layer3.2.bn3.running_mean", "Backbone.model.layer3.2.bn3.running_var", "Backbone.model.layer3.2.bn3.num_batches_tracked", "Backbone.model.layer3.3.conv1.weight", "Backbone.model.layer3.3.bn1.weight", "Backbone.model.layer3.3.bn1.bias", "Backbone.model.layer3.3.bn1.running_mean", "Backbone.model.layer3.3.bn1.running_var", "Backbone.model.layer3.3.bn1.num_batches_tracked", "Backbone.model.layer3.3.conv2.weight", "Backbone.model.layer3.3.bn2.weight", "Backbone.model.layer3.3.bn2.bias", "Backbone.model.layer3.3.bn2.running_mean", "Backbone.model.layer3.3.bn2.running_var", "Backbone.model.layer3.3.bn2.num_batches_tracked", "Backbone.model.layer3.3.conv3.weight", "Backbone.model.layer3.3.bn3.weight", "Backbone.model.layer3.3.bn3.bias", "Backbone.model.layer3.3.bn3.running_mean", "Backbone.model.layer3.3.bn3.running_var", "Backbone.model.layer3.3.bn3.num_batches_tracked", "Backbone.model.layer3.4.conv1.weight", "Backbone.model.layer3.4.bn1.weight", "Backbone.model.layer3.4.bn1.bias", "Backbone.model.layer3.4.bn1.running_mean", "Backbone.model.layer3.4.bn1.running_var", "Backbone.model.layer3.4.bn1.num_batches_tracked", "Backbone.model.layer3.4.conv2.weight", "Backbone.model.layer3.4.bn2.weight", "Backbone.model.layer3.4.bn2.bias", "Backbone.model.layer3.4.bn2.running_mean", "Backbone.model.layer3.4.bn2.running_var", "Backbone.model.layer3.4.bn2.num_batches_tracked", "Backbone.model.layer3.4.conv3.weight", "Backbone.model.layer3.4.bn3.weight", "Backbone.model.layer3.4.bn3.bias", "Backbone.model.layer3.4.bn3.running_mean", "Backbone.model.layer3.4.bn3.running_var", "Backbone.model.layer3.4.bn3.num_batches_tracked", "Backbone.model.layer3.5.conv1.weight", "Backbone.model.layer3.5.bn1.weight", "Backbone.model.layer3.5.bn1.bias", "Backbone.model.layer3.5.bn1.running_mean", "Backbone.model.layer3.5.bn1.running_var", "Backbone.model.layer3.5.bn1.num_batches_tracked", "Backbone.model.layer3.5.conv2.weight", "Backbone.model.layer3.5.bn2.weight", "Backbone.model.layer3.5.bn2.bias", "Backbone.model.layer3.5.bn2.running_mean", "Backbone.model.layer3.5.bn2.running_var", "Backbone.model.layer3.5.bn2.num_batches_tracked", "Backbone.model.layer3.5.conv3.weight", "Backbone.model.layer3.5.bn3.weight", "Backbone.model.layer3.5.bn3.bias", "Backbone.model.layer3.5.bn3.running_mean", "Backbone.model.layer3.5.bn3.running_var", "Backbone.model.layer3.5.bn3.num_batches_tracked", "Backbone.model.layer4.0.conv1.weight", "Backbone.model.layer4.0.bn1.weight", "Backbone.model.layer4.0.bn1.bias", "Backbone.model.layer4.0.bn1.running_mean", "Backbone.model.layer4.0.bn1.running_var", "Backbone.model.layer4.0.bn1.num_batches_tracked", "Backbone.model.layer4.0.conv2.weight", "Backbone.model.layer4.0.bn2.weight", "Backbone.model.layer4.0.bn2.bias", "Backbone.model.layer4.0.bn2.running_mean", "Backbone.model.layer4.0.bn2.running_var", "Backbone.model.layer4.0.bn2.num_batches_tracked", "Backbone.model.layer4.0.conv3.weight", "Backbone.model.layer4.0.bn3.weight", "Backbone.model.layer4.0.bn3.bias", "Backbone.model.layer4.0.bn3.running_mean", "Backbone.model.layer4.0.bn3.running_var", "Backbone.model.layer4.0.bn3.num_batches_tracked", "Backbone.model.layer4.0.downsample.0.weight", "Backbone.model.layer4.0.downsample.1.weight", "Backbone.model.layer4.0.downsample.1.bias", "Backbone.model.layer4.0.downsample.1.running_mean", "Backbone.model.layer4.0.downsample.1.running_var", "Backbone.model.layer4.0.downsample.1.num_batches_tracked", "Backbone.model.layer4.1.conv1.weight", "Backbone.model.layer4.1.bn1.weight", "Backbone.model.layer4.1.bn1.bias", "Backbone.model.layer4.1.bn1.running_mean", "Backbone.model.layer4.1.bn1.running_var", "Backbone.model.layer4.1.bn1.num_batches_tracked", "Backbone.model.layer4.1.conv2.weight", "Backbone.model.layer4.1.bn2.weight", "Backbone.model.layer4.1.bn2.bias", "Backbone.model.layer4.1.bn2.running_mean", "Backbone.model.layer4.1.bn2.running_var", "Backbone.model.layer4.1.bn2.num_batches_tracked", "Backbone.model.layer4.1.conv3.weight", "Backbone.model.layer4.1.bn3.weight", "Backbone.model.layer4.1.bn3.bias", "Backbone.model.layer4.1.bn3.running_mean", "Backbone.model.layer4.1.bn3.running_var", "Backbone.model.layer4.1.bn3.num_batches_tracked", "Backbone.model.layer4.2.conv1.weight", "Backbone.model.layer4.2.bn1.weight", "Backbone.model.layer4.2.bn1.bias", "Backbone.model.layer4.2.bn1.running_mean", "Backbone.model.layer4.2.bn1.running_var", "Backbone.model.layer4.2.bn1.num_batches_tracked", "Backbone.model.layer4.2.conv2.weight", "Backbone.model.layer4.2.bn2.weight", "Backbone.model.layer4.2.bn2.bias", "Backbone.model.layer4.2.bn2.running_mean", "Backbone.model.layer4.2.bn2.running_var", "Backbone.model.layer4.2.bn2.num_batches_tracked", "Backbone.model.layer4.2.conv3.weight", "Backbone.model.layer4.2.bn3.weight", "Backbone.model.layer4.2.bn3.bias", "Backbone.model.layer4.2.bn3.running_mean", "Backbone.model.layer4.2.bn3.running_var", "Backbone.model.layer4.2.bn3.num_batches_tracked", "Backbone.model.fc.weight", "Backbone.model.fc.bias", "regressionModel.conv1.weight", "regressionModel.conv1.bias", "regressionModel.bn1.weight", "regressionModel.bn1.bias", "regressionModel.bn1.running_mean", "regressionModel.bn1.running_var", "regressionModel.bn1.num_batches_tracked", "regressionModel.conv2.weight", "regressionModel.conv2.bias", "regressionModel.bn2.weight", "regressionModel.bn2.bias", "regressionModel.bn2.running_mean", "regressionModel.bn2.running_var", "regressionModel.bn2.num_batches_tracked", "regressionModel.conv3.weight", "regressionModel.conv3.bias", "regressionModel.bn3.weight", "regressionModel.bn3.bias", "regressionModel.bn3.running_mean", "regressionModel.bn3.running_var", "regressionModel.bn3.num_batches_tracked", "regressionModel.conv4.weight", "regressionModel.conv4.bias", "regressionModel.bn4.weight", "regressionModel.bn4.bias", "regressionModel.bn4.running_mean", "regressionModel.bn4.running_var", "regressionModel.bn4.num_batches_tracked", "regressionModel.output.weight", "regressionModel.output.bias", "classificationModel.conv1.weight", "classificationModel.conv1.bias", "classificationModel.bn1.weight", "classificationModel.bn1.bias", "classificationModel.bn1.running_mean", "classificationModel.bn1.running_var", "classificationModel.bn1.num_batches_tracked", "classificationModel.conv2.weight", "classificationModel.conv2.bias", "classificationModel.bn2.weight", "classificationModel.bn2.bias", "classificationModel.bn2.running_mean", "classificationModel.bn2.running_var", "classificationModel.bn2.num_batches_tracked", "classificationModel.conv3.weight", "classificationModel.conv3.bias", "classificationModel.bn3.weight", "classificationModel.bn3.bias", "classificationModel.bn3.running_mean", "classificationModel.bn3.running_var", "classificationModel.bn3.num_batches_tracked", "classificationModel.conv4.weight", "classificationModel.conv4.bias", "classificationModel.bn4.weight", "classificationModel.bn4.bias", "classificationModel.bn4.running_mean", "classificationModel.bn4.running_var", "classificationModel.bn4.num_batches_tracked", "classificationModel.output.weight", "classificationModel.output.bias", "DepthRegressionModel.conv1.weight", "DepthRegressionModel.conv1.bias", "DepthRegressionModel.bn1.weight", "DepthRegressionModel.bn1.bias", "DepthRegressionModel.bn1.running_mean", "DepthRegressionModel.bn1.running_var", "DepthRegressionModel.bn1.num_batches_tracked", "DepthRegressionModel.conv2.weight", "DepthRegressionModel.conv2.bias", "DepthRegressionModel.bn2.weight", "DepthRegressionModel.bn2.bias", "DepthRegressionModel.bn2.running_mean", "DepthRegressionModel.bn2.running_var", "DepthRegressionModel.bn2.num_batches_tracked", "DepthRegressionModel.conv3.weight", "DepthRegressionModel.conv3.bias", "DepthRegressionModel.bn3.weight", "DepthRegressionModel.bn3.bias", "DepthRegressionModel.bn3.running_mean", "DepthRegressionModel.bn3.running_var", "DepthRegressionModel.bn3.num_batches_tracked", "DepthRegressionModel.conv4.weight", "DepthRegressionModel.conv4.bias", "DepthRegressionModel.bn4.weight", "DepthRegressionModel.bn4.bias", "DepthRegressionModel.bn4.running_mean", "DepthRegressionModel.bn4.running_var", "DepthRegressionModel.bn4.num_batches_tracked", "DepthRegressionModel.output.weight", "DepthRegressionModel.output.bias".

[ccchen091]

您好，我还想问一下我按照A2J的方式处理好NYU数据集以后，运行代码出现这个通道数不匹配问题：RuntimeError: Given groups=1, weight of size [64, 3, 7, 7], expected input[4, 1, 176, 176] to have 3 channels, but got 1 channels instead 请问一下这个是为什么呢？并且您有什么好的建议吗？

[ChanglongJiangGit]

您好，model_41_wetD_0.0001_depFact_0.5_RegFact_3_rndShft_5.pth这个文件忘记公布了，稍后会上传，请您留意！您也可以先自行训练，训练出的结果应该不会有多少差别。

对于第二个问题，请问您现在解决了吗？没解决的话是否可以告诉我出现问题的代码位置？谢谢！

[ccchen091]

您好，model_41_wetD_0.0001_depFact_0.5_RegFact_3_rndShft_5.pth这个文件忘记公布了，稍后会上传，请您留意！您也可以先自行训练，训练出的结果应该不会有多少差别。

对于第二个问题，请问您现在解决了吗？没解决的话是否可以告诉我出现问题的代码位置？谢谢！

谢谢您的回复! 目前我们的任务没用上NYU数据集了，所以就没去解决这个问题。

[ccchen091]

您好！我想请问您一个问题，我在您的模型基础上训练了其他数据集（assemblyands），基本上就只改动了数据读取dataset，但是训练完以后，test的指标特别低，mpjpe94，请问一下这个是模型和数据集不适配吗？还是我有什么其他地方没注意的？

[zzsyid]

您好，我在NYU上训练时也出现RuntimeError: Given groups=1, weight of size [64, 3, 7, 7], expected input[4, 1, 176, 176] to have 3 channels, but got 1 channels instead，具体问题出现在 File "/A2J-Transformer-main/model.py", line 285, in forward features, pos = self.backbone(samples)File "/A2J-Transformer-main/dab_deformable_detr/backbone.py", line 90, in forward xs = self.body(tensor_list.tensors) 是否是NYU数据集还需要额外的预处理或者更改backbone的维度参数呢

[ccchen091]

您好 我想再请教您一个问题 我将您的模型在Interhand2.6M数据集上训练好以后，我拍了一张相同视角的图片进去可视化看效果，按理说我自己拍的图片是不需要内外参数就可以得到不错的可视化效果的是吗？但是事实是我的可视化效果特别差，完全对应不上，请问一下这是什么问题呢？

[ChanglongJiangGit]

您好！我想请问您一个问题，我在您的模型基础上训练了其他数据集（assemblyands），基本上就只改动了数据读取dataset，但是训练完以后，test的指标特别低，mpjpe94，请问一下这个是模型和数据集不适配吗？还是我有什么其他地方没注意的？

这个具体不清楚，建议可视化检查一下代码和数据是否匹配等等

[ChanglongJiangGit]

您好 我想再请教您一个问题 我将您的模型在Interhand2.6M数据集上训练好以后，我拍了一张相同视角的图片进去可视化看效果，按理说我自己拍的图片是不需要内外参数就可以得到不错的可视化效果的是吗？但是事实是我的可视化效果特别差，完全对应不上，请问一下这是什么问题呢？

您好呀，请问一下您可视化的是模型输出的2.5D结果吗？如果模型输出的2.5D结果很差，那这个可能是由于拍摄的图片跟InterHand数据集差异过大导致的；如果模型预测的2.5D结果是正常的，但是转换到相机空间下不正常，那可能是由于相机参数有问题导致的。