sove45
sove45
I used train3d.py for training, saved the weights after training, and then loaded them. We found that regardless of which parameters I set as trainable, the network converged based on...
Have you conducted training with masks as prompts in videos? I noticed that your paper mentions related training.
关于mask prompt的在3D数据集上/视频序列的训练方法Training Methods for Mask Prompting on 3D Datasets / Video Sequences
我注意到在您的代码中也存在着 I noticed that your code also contains  在训练3d数据集的相关代码，可是当我使用这个函数来尝试利用第一帧的mask作为prompt来进行训练的时候，发现出现第一帧没有梯度的报错，我通过debug发现是在代码中，对于有mask prompt的帧直接输出输入mask导致的， In the code related to training on 3D datasets, when I tried to use the mask from the first...
In the ScanNet 2D segmentation annotations, there are two versions of masks—instance vs. instance-filt and label vs. label-filt. For your experiments, which version did you use as the ground truth...
In the ScanNet 2D segmentation annotations, there are two versions of masks—instance vs. instance-filt and label vs. label-filt. For your experiments, which version did you use as the ground truth...
In the ScanNet 2D segmentation annotations, there are two versions of masks—instance vs. instance-filt and label vs. label-filt. For your experiments, which version did you use as the ground truth...