Image Recognition using CIFAR-10 dataset
How to use this repository?
- load_cifar.py- It is the model training file which consists of dataset split, normalization, developing neural network, training and storing the structure as well as the weights in json format.
This is the neural network model that is built.
![Model summary](https://github.com/sidvsukhi/Image-Recognition-CIFAR-10/blob/master/summary.PNG)
2. use_cifar.py- It is the file that uses the stored neural network structure and weights to predict.
![Test image](https://github.com/sidvsukhi/Image-Recognition-CIFAR-10/blob/master/frog.jpg)
3. model_structure.json- It is the stored json file with the trained neural network structure
4. model_weights.h5- It has the stored weights of the model.
[Medium article](https://medium.com/@siddhantsukhatankar/hello-world-in-image-recognition-136eba581464)
#################
pip3 install -r requirements.txt
export MLFLOW_TRACKING_URI="sqlite:///mlruns.db"
python3 model_service.py
mlflow models serve -m models:/SentimentAnalysis/latest -p 5000 --no-conda
curl http://127.0.0.1:5000/invocations -H 'Content-Type: application/json' -d '{"instances":[ "It is a bad day"]}' or ###################