xxxhy
xxxhy
But actually I found here, when [try to get mesh gt's path](https://github.com/GAP-LAB-CUHK-SZ/InstPIFu/blob/main/evaluate_object_reconstruction.py#L135C15-L135C15), there will be a lot of mesh gts couldn't be found. Is it correct? And also, I reproduce...
I see. Thank you for your patience and wait for the conclusion!
Hi, is the quantitative results evaluated on the "wrong" watertight meshes or the updated ones?
Also, why desk is None? Have you removed all the desk samples?
> @xXuHaiyang The desk is no longer None currently and I update it in the readme file. The quantitative results is run on the updated file, which is zip as...
@UncleMEDM Hi, I wonder what do "split" and "split-filter" actually mean? In Object Reconstrcution Task configs, I found the data path in [train.json](https://github.com/GAP-LAB-CUHK-SZ/InstPIFu/blob/main/configs/train_instPIFu.yaml) is "split-filter", while in [test.json](https://github.com/GAP-LAB-CUHK-SZ/InstPIFu/blob/main/configs/test_instPIFu.yaml) is "split"...
冒昧回答，事实上在这里 [unmix_x_rec](https://github.com/Sense-X/MixMIM/blob/master/models_mixmim.py#L421) ，由于 `mask_ratio` 为 50%，因此 `x1_rec` 和 `x2_rec` 是“互补”的。这两份“互补”的 tokens 再加起来，就是一张“完全”图像的 reconstruction 结果，对应的 target 也就是单张图像。