search

littlebeen / DDPM-Enhancement-for-Cloud-Removal

A cloud removal dataset and a diffusion-based cloud removal method
29 stars 1 forks source link

weight of the pretrained model #4

Open mboumahdi opened 3 months ago

mboumahdi commented 3 months ago

Hello, I'm trying to run the pretrained model you previously share with me in this link https://huggingface.co/littlebeen/DE/tree/main

but for the model mn im calling this class model = MPRNet().to('cuda') with this model _RICE2mnt/ema_0.9999010000.pt

and mdsat model with model = Generate_quarter(in_channels=4,height=3,width=6,num_dense_layer=4,growth_rate=16).to('cuda') and file _RICE2mdsat/ema_0.9999010000.pt

But i'm getting mismatched weight keys in both, i don't know what could be the reason

Thank you

littlebeen commented 2 months ago

I have update the code structure and the pretrain model. Please refer to readme to make better preparation before training and testing.

littlebeen commented 2 months ago

Download the pretain e2e model (mdsa/mn as your need) and put it into guided_diffusion/couldnet/mdsa/pretrain or guided_diffusion/couldnet/mn/pretrain. The model could be found at https://pan.baidu.com/s/1lyKRG67AxM5SZEj3mhTKWA code:bean

You could choose the appropriate pretrain model as your need (Which model You want to train or test based on? memorynet or mdsa?). Must be done before training and testing

littlebeen commented 2 months ago

Hope it could help u!

mboumahdi commented 2 months ago

Thank you for the update, but with my location outside china I can't access to Baidu link