FruityFit

CH2-PS327

CH2-PS327 Contributors

Name	Bangkit-ID	Learning Path
Evander Gabriel	M239BSY0797	Machine Learning
Christopher Ade Wiyanto	M239BSY1097	Machine Learning
Sonya Oktavia	M200BSX1152	Machine Learning
Danar Hadi Bachtiar	C253BSY3294	Cloud Computing
Muhamad Arya Al Ghifari Wibowo	C134BSY3568	Cloud Computing
Melisa Wijaya	A239BSX2324	Mobile Development
Adam Rayhant Laksono	A009BSY2477	Mobile Development

Machine Learning

1. Fruit Classifications

Steps

1. Collect dataset
2. Make 2 folders for test and train
3. Make folders in each folders as much classes and name it.
4. Use ImageDataGenerator to augment the images dataset
5. Tensorflow to build the ML model using CNN
6. Make the REST API using Flask

2. Fruit Freshness Classifications

Steps

1. Collect dataset
2. Make 2 folders for test and train
3. Make 2 folders in each folder because we want binary and name it True or False.
4. Use ImageDataGenerator to augment the images dataset
5. Tensorflow to build the ML model using CNN
6. Save to SavedModel then convert to tensorflow.lite

3. Chatbot

Botie employs text classification to predict future events and utilizes various Google products, such as Generative AI, Palm API, TensorFlow, and a custom search API. In addition to typical chatbot functions, Botie examines user history to provide personalized juice recommendations based on individual preferences and nutritional requirements

Steps

1. Gather data
2. Create separate directories for testing and training
3. Generate JSON for the data
4. Apply NLTK preprocessing for text data
5. Utilize NLTK for data augmentation and manipulation
6. Saved it to class blueprint for advanced deployment
7. Construct a class blueprint for 2 generative ai and web scraping
8. Make one class to combine 3 class before for deployment
9. Save as a SavedModel and convert to H5
10. deployed in flask server

Result