Skip to content

Meta learning for few-shot learning task on miniImagenet, using ResNet-18 as a feature extractor. Project for CMPT 726

Notifications You must be signed in to change notification settings

sirandou/meta-learning-few-shot-classification

Repository files navigation

ML-course project

Meta learning for few-shot learning task, using ResNet-18

based on https://github.com/floodsung/LearningToCompare_FSL

Download images for mini-image-net from: https://drive.google.com/file/d/0B3Irx3uQNoBMQ1FlNXJsZUdYWEE/view and put it in datas/miniImagenet directory and run proc_images.py

Put all the files in your Google drive under the directory 'ML project'.

You can run both notebooks step by step to see how both models (baseline and ResNet-18 feature extraction) perform. For evaluation, the trained networks are saved in the models directory. Enjoy!

About

Meta learning for few-shot learning task on miniImagenet, using ResNet-18 as a feature extractor. Project for CMPT 726

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published