Validation of finetuning with Resnet18 using Pytorch on Caltech101.
Pretrained.py is the net fintuned based on Resnet18 and scratch.py serves as control.
Before model training, you have to download Caltech101 and run data.py for data preprocessing first.
CAM.py is added for visualization of experiment results. The algorithm is based on Learning Deep Features for Discriminative Localization.
Michi-3000/FinetuningBasedonResnet18
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|