ValueError: GraphDef cannot be larger than 2GB.

[Kaushalya]

Thank you very much for sharing the code publicly.

I tweaked the code to run it on GPU without requiring much memory. (can send a PR) When I ran calculate_hessian_inverse.py I get the following error

ValueError: GraphDef cannot be larger than 2GB.

I could create the hessian inverse matrices of ResNet-50 by running calculate_hessian_inverse.py several times for the remaining layers.

Have you encountered this issue? I could not figure out why the graph keeps growing.

[csyhhu]

Hi @Kaushalya ,

Thanks for using our codes. For the memory issue, I suggest that you can increase the stride in tf.extract_images_pathes(), so that the number of extracted patches will be decrease.

BTW, we are now refining the codes and transfer it to PyTorch, codes will be clearer and easier to operate, and decrease the demand for memory. The new codes will be published in days. I suggest you to wait then use the new codes.