This repository contains six machine learning projects and modules that demonstrate the application of machine learning techniques to solve real-world problems.

This project demonstrates a machine learning application for classifying food images using transfer learning and fine-tuning techniques with TensorFlow. The project aims to achieve high accuracy in classifying food images into 101 different classes, utilizing a large dataset of over 75,000 images from the Food101 dataset. The project focuses on building a robust model that can accurately classify food images into their respective categories, with a particular emphasis on precision, recall, and f1 scores.

This project focuses on building a machine learning model to classify medical abstract senstences into different sentence classes using natural language processing (NLP) models. The project utilizes NLP techniques to analyze the text data and build a classification model using 200,000+ labelled randomized abstracts from the PubMed 200k RCT dataset. The goal is to accurately classify the sentences into their respective classes, which can be useful for various applications in the medical field, such as information retrieval and summarization.

This project focuses on building a machine learning model to predict future values in a time series dataset. The goal is to forecast the price of a specific asset based on historical data. The project utilizes recurrent neural networks (RNNs) and long short-term memory (LSTM) networks to capture temporal dependencies in the data. The project includes steps such as data pre-processing, feature engineering, model training, and evaluation. The model is trained on a dataset of historical price data and aims to achieve high accuracy in forecasting future prices.

This module provides an overview of pre-trained ImageNet models in TensorFlow and Keras. It covers the use of a pretrained model for image classification tasks, including loading a pre-trained model, preprocessing input images, making predictions, and decoding the results.

This module focuses on using feedforward multilayer networks for hand-written digit classification tasks in Keras. It covers the steps in data pre-processing, model architecture inference, evaluating its performance, and making predictions.

This module presents a traffic sign image classification using transfer learning and fine-tuning techniques with TensorFlow and Keras. It covers the steps in data pre-processing, model inference, freezing and compiling the model, evaluating its performance, and making predictions on test dataset.