Not torchvision model case

[AlexanderSif]

Hello,

I'm trying to use post-training quantization on a CNN with brevitas Xilinx, but I'm having trouble getting it to work. I have read the README in the ptq folder and realised that ptq_evaluate.py seems to be the solution. However, it appears to accept only a specific type of models (torchvision.models). I'd like to apply it to one of my own models. Have I missed any parameters or is it not possible to use it with models other than torchvision models?

[Giuseppe5]

The script itself is meant only for torchvision models.
You can try modifying it and pass your own model, but results might vary based on the model structure.

Let us know if you encounter any problem.