-
Hi, I'm trying to run the training scripts to replicate the results but DiffHOI-L does not work.
In the network class, at initialization time `self.class_embed = nn.ModuleList(class_embed_layerlist)…
-
如何在YOLO的backbone基础上同时加入Efficientnet,使其在neck部分进行融合即可。大佬可以给个样吗?有偿
-
A set of backbone and decoder models must be supported in terratorch, as U-Nets and ASPP (Atrous Spatial Pyramid Pooling).
-
Hello, Thanks for your amazing work. I was wondering if you had performed ablation studies on using backbone alignment? I assume the performance did not improve after using all four alignment strategi…
-
-
Hello,
Could you provide the vision transformer backbone used for the model?
I am using dino's vision_transformer.py code for a vit -giant (https://github.com/facebookresearch/dino/blob/main/visi…
-
Hi! Can you provide codes to reproduce MACs in Table 6(b) in your paper?
I calculate MACs on multiHMR_896_L using fvcore (https://github.com/facebookresearch/fvcore), but I get different outputs from…
-
Great package that is already the fastest python implementation as far as I know. Still, are there plans to use C code to speed up computations, similar to how Levenstheins python port does? Would mak…
-
请问下如何冻结代码库里的hivit呢?像vit和resnet可以使用frozen_stages参数结合_freeze_stages函数来进行冻结,hivit该怎么操作?
-
作者您好!代码默认的配置是'backbone_name': 'ViT-L-14-336', 'feature_layers': [5, 11, 17, 23], 'r_list': [1, 3, 5], 我将backbone_name改成ViT-B-16后运行报错,我推测是要修改相应的feature_layers或者r_list,可以告诉我怎样修改吗?谢谢!