How can i know the architecture of pre-trained model in Pytorch?

Question

I have downloaded this pre-trained model "model_ir_se50.pth" for face recognition. It is giving very good results. How can i know its architecture?

Does this answer your question? [Model summary in pytorch](https://stackoverflow.com/questions/42480111/model-summary-in-pytorch) — Dishin H Goyani, Aug 19 '20 at 08:54

score 0 · Answer 1 · answered Aug 19 '20 at 08:27

0

simple print(net) #replace net with the network variable will work as per your need.

answered Aug 19 '20 at 08:27

Mughees

you mean like this : `import torch pretrained = torch.load("model_ir_se50.pth") print(pretrained.net()) ` – kritika Garg Aug 19 '20 at 08:33
`print(pretrained)` – Mughees Aug 19 '20 at 08:36
Its output is very big..but you can get an idea from a part of it i am pasting. OrderedDict([('input_layer.0.weight', tensor([[[[ 5.0732e-02, 6.5002e-03, 1.9098e-02],.... – kritika Garg Aug 19 '20 at 08:43
can you edit your post and add the code which you are using to get the results. – Mughees Aug 19 '20 at 09:02

1 Answers1