Hello, thank you for your great work.
I meet some problems when training MindEye. I downloaded the pretrained model for subject1, and found the reconstruction performance is good as expected.
However, when I attempted to train a new model with Train_MindEye.py, I found the reconstuction results all poor. Especially, when using model with pretrained weights, the test loss is much higher than model with random Initialization. and it seem to be wrong when the optimizers load_state_dict of your public pretrained weights.
So i wanna to know whether there are some errors in training script or rather, I didn't do it right.
Thanks again.
Hello, thank you for your great work. I meet some problems when training MindEye. I downloaded the pretrained model for subject1, and found the reconstruction performance is good as expected. However, when I attempted to train a new model with Train_MindEye.py, I found the reconstuction results all poor. Especially, when using model with pretrained weights, the test loss is much higher than model with random Initialization. and it seem to be wrong when the optimizers load_state_dict of your public pretrained weights. So i wanna to know whether there are some errors in training script or rather, I didn't do it right. Thanks again.