Mamba.__init__() got an unexpected keyword argument layer_idx

[Shadow581]

Hello, I has followed your instruction to install env, but it shows errors.

link: SkyworkAI/PointCloudMamba#9

This problem bothered me for a long time.

[xiaoyao3302]

Hi, thanks for your interest in our work. Did you come across this problem when running our code?

[Shadow581]

Thanks your reply.

I followed your instruction to install. (except python, torch and cuda version)

Execute your code, but it shows errors about knn_cuda Lib.

My cuda is available:

[xiaoyao3302]

Is CUDA version right? Does it match the knn dependencies? I remembered that installing knn and FPS requires a determined CUDA version. BTW what does GPT say?

[Shadow581]

Is CUDA version right? Does it match the knn dependencies? I remembered that installing knn and FPS requires a determined CUDA version. BTW what does GPT say?

Is this the wrong match? I am using a cloud platform and I didn't configure the torch and cuda versions

I only have one cuda version, cuda-11.8.

I following advices of GPT, but it still shows errors.

[Shadow581]

Additionally, I run my own code on the same env, it doesn't show errors. It can get knn data.

[Shadow581]

Is CUDA version right? Does it match the knn dependencies? I remembered that installing knn and FPS requires a determined CUDA version. BTW what does GPT say?

Hello, I sovle the problem. Set CUDA_VISIBLE_DEVICES=0 in run.sh file.

[Shadow581]

Hello, everyone, I have sovled the problem stated in the title.
I configured Mamba to just copy the models/block.py/Block code into my code, but this gave me an error(such as title). You need to copy all the functions about wrapping Mamba code into your code (up to the MixerModel class).

[xiaoyao3302]

Cool