search

BehnoodRasti / HySUPP

An Open-Source Hyperspectral Unmixing Python Package
MIT License
39 stars 5 forks source link

fig error #2

Open smlba opened 10 months ago

smlba commented 10 months ago
PixPin_2023-11-30_20-34-53 PixPin_2023-11-30_20-35-11 PixPin_2023-11-30_20-41-45 PixPin_2023-11-30_20-35-34

First of all, I want to say thank you for your amazing work on HySUPP. I really appreciate your efforts and contributions to the field of hyperspectral image unmixing. However, I have a question about running the code. As you can see in the first image, I ran the sample code in the command terminal, and after a long time waiting (fig2), I got the solution (fig3). However, I could not see the abundance image（fig4). Maybe I had a wrong configuration, so I have to describe what I did after I cloned the GitHub repository. First, I used a conda virtual Python environment to install HySUPP. Then, I installed the required Python packages and installed spams-bin instead of spams. At the same time, I installed the module matlab.engine for Python, and adjusted the MATLAB_root path to the installation location of MATLAB on my computer. That is all. In the end, I ran the code.

I hope you can help me solve this problem and explain why I could not see the abundance image. Thank you very much for your time and attention. I look forward to hearing from you soon.

hsiaotung-tan commented 7 months ago

Hi, can I know which logger is inside your config.yaml? I got an error : AttributeError: 'DefaultLogger' object has no attribute 'log_artifact'

azouaoui-cv commented 7 months ago

Hi @smlba, I apologize for not having come sooner to the rescue. It seems you used SUnCNN with 20000 iterations on the DC1 image, am I correct? First of all, did you input any SNR (e.g. noise.SNR=30)? It seems that you ran the program with the CPU and not GPU, hence it was very slow (3hours+). Can you confirm? To give you an idea, it takes around 3 mins to run 20000 iterations on the DC1 image with SUnCNN on a NVIDIA RTX 2080 Ti. If I'm not mistaken, you should find the estimated abundances in the file logs/1/artifacts/Estimate/estimates.mat. Let me know if that helps!