L

Check whether your GraphDef-interpreting binary is up to date with your GraphDef-generating binary.

1

my producer_version = default_graph.graph_def_versions.producer is 24 so i need ti download the ['net-lin_alex_v0.1_24.pb', 'net-lin_alex_v0.1.pb'], but there is not net-lin_alex_v0.1_24.pb. how can i get it? when i run the test_network.py, there...

why not do_augmentation for kp_3d？

threedpw_utils.py 疑惑

1

threedpw_utils.py 采用默认的threedpw_utils代码处理3dpw的数据时候，采用下面的代码获得vertices #output = smpl(betas=shape, body_pose=pose[:,3:], global_orient=pose[:,:3], transl=trans)#原始代码 发现这个vertices与spin网络获得cam并不一致，原因是这里传入了3dpw的trans参数， 而vibe模型参数的中，网络训练的输出，并没有传入trans参数，代码如下，不知道是不是弄错了， 或者在代码里面哪里做了转换，望指教，谢谢 pred_output = self.smpl( betas=pred_shape, body_pose=pred_rotmat[:, 1:], global_orient=pred_rotmat[:, 0].unsqueeze(1), pose2rot=False )

IOError: Failed to interpret file '/home/Faster-RCNN_TF/data/pretrain_model/VGG_imagenet.npy' as a pickle

1

File "/home/Faster-RCNN_TF/tools/../lib/networks/networks.py", line 51, in load data_dict = np.load(data_path).item()

some questions about the train code

10

1 romp/lib/dataset/image_base.py 你好 请问下，valid_masks = np.zeros((self.max_person, 6), dtype=np.bool) 第二维度一共是6种类型，不知道523行这里为什么有判断，并且下标是6，超过了valid_masks的范围。   https://github.com/Arthur151/ROMP/blob/ee5e2f21f35a1072327a11ecd4a36c0c64d805e1/romp/lib/dataset/image_base.py#:~:text=if%20r%5B%27valid_masks%27%5D%5B0%2C0%2C6%5D%3A

Have you used 3d loss for fitting about simplify-X?

25

Thanks a lot for sharing the Simplify-X fits for H36M. Did you use 3d loss when fitting? I found that his side view is slanted, indicating that the depth is...

the paper have Mentioned “train a copy of our model (“Ours-png”) using 8-bit processed RGB images to evaluate the benefits of having real raw sensor data”

数据细节问题

11

你好，很感谢提供了3dpw和hm36的标签，这里有几个疑惑。 关于pw3d：  问题1：为什么使用的betas是annots_hmrvideo的而不是pw3d原始提供的beta； 问题2：label['cam_rotation_matrix'] = raw_labels['cam_poses'][j,:3,:3] ，看你的centerHMR，这个相机外参矩阵应该是乘以到pose的global_rotation。在进行最终的监督时，不要单独监督这个相机外参。但是看你最终返回的dataset，好像没有使用这个相机外参。  所以这里的pose，其实没有对global_rotation做额外的处理？

hdrnet_ops.so 求提供

6

hdrnet_ops.so 一座大山

why there are No module named caffe? but i have make in caffe with [100%] Built target cmp_trie

2

$ make [ 50%] Built target trie_py [100%] Built target cmp_trie when python demo.py Traceback (most recent call last): File "demo.py", line 10, in import caffe ImportError: No module named...