I am going to implement Resnet model using transfer learning, finetuning the model to predict cats and dogs... then convert the model to ONNX and test it again using onnxruntime... and at last stream it using streamlit...
you can find the data i'm using in this mini project here: "https://www.kaggle.com/competitions/dogs-vs-cats/data"
1- Download cats and dogs dataset ✔️
2- Prepare it to make dataset class ✔️
3- Build a dataset class and show a randome image using dataloader ✔️
4- Understand the architecture of ResNet model ✔️
5- Implement pretrained ResNet model and finetune it ✔️
6- Train the model to achive accuracy of more than 90% ✔️
7- Save and Test the best model ✔️
8- Implemetnation of ONNX ✔️
9- Run the project in streamlit ✔️
you can see the final result on ui_streamlit.py. run it using this line of code:
streamlit run ui_streamlit.py
(no need to download the dataset)