Skip to content

Repository files navigation

#The modified data set contains 5000 apple pictures, including 2000 rotten apples and 2000 fresh apples,Divide training set and test set by 4:1 #Choose ResNet18 for migration learning, and only fine-tune its fully connected layer (last layer), where batch_size is 4 and epoch is 6 #We get a model that is 97% accurate on the test set #if you want the apple dataset,welcome contact with me.

#email:756534202@qq.com

About

它可以识别苹果是否腐烂,并给出详细的新鲜度分析

Topics

Resources

Stars

10 stars

Watchers

1 watching

Forks

Releases

Packages

Contributors

Languages