vitae-transformer/mtp

The official repo for "MTP: Advancing Remote Sensing Foundation Model via Multi-Task Pretraining"

mtp's Issues

运用

你好，我想请问一下我可以直接用你们已发布的模型，来估计自己数据集图像中目标数量吗，怎样做更加方便呢？谢谢！

我想请问一下你们tools/train.py可以公开一下吗

Look forward to the release of the training code

Quite leading work, the release of the training code is anticipated.

测试

请问应该如何用你训练好的模型做测试呢，没有找到tool/test.py，在windows系统下可以跑吗

Installing the repo and using MMSeg

Dear authors, thank you for your work and for open-sourcing it. I have a question about using this code.

I am not familiar with MMSeg, so this is probably a stupid question. I have installed it using pip but this does not install the tools folder that is needed to run the commands in the README. Should I have instead used the MMSeg repository instead of the pip package ? If yes, how do I use your repo with respect to the MMSeg repo ? It seems that MMSeg does not work like the libraries I am used to, so I am a bit confused.

Thank you in advance and good luck with your future works.

🐢🐢, I am so psyched about this work!!

How to inference on custom dataset?

Hello, thanks for your amazing job.
I want to use Semantic Segmentation model finetuned on LoveDA, but it seems that there is no script for inference on custom dataset provided here.
How should I do it?

Multichannel images

Hello,

It seems that the examples focus on 3 channel images.

Is semantic segmentation expected to work on multichannel images, and is there an example on that?

I know that mmsegmentation has LoadSingleRSImageFromFile, but I wonder if there would be any specific limitation associated with MTP? How to handle pre-trained weights?

Thanks!