CIFAR-10 Image Classification & AI Agent with RAG

This repository contains implementations of two tasks:

1. CIFAR-10 Image Classification using Deep Learning
2. AI Agent with RAG and Function Calling

Task 1: CIFAR-10 Image Classification

Objective

Train a deep learning neural network to classify images using the CIFAR-10 dataset, which contains 60,000 labeled images of 10 different classes.

Dataset

• CIFAR-10 Dataset: Imported from TensorFlow/Keras datasets.

Features Implemented

• Data Preprocessing:
  • Loaded the dataset directly from TensorFlow/Keras.
  • Normalized pixel values to the range [0,1].
  • Split the dataset into training and test sets.
• Built a CNN Model using:
  • Conv2D, MaxPooling2D, Flatten, Dense, Dropout, and Batch Normalization layers.
• Training & Validation:
  • Applied proper train-test split.
  • Used early stopping to prevent overfitting.
  • Tuned hyperparameters such as learning rate and batch size for better performance.
• Performance Evaluation:
  • Computed Accuracy, Precision, Recall, and F1-score.
  • Plotted training and validation accuracy/loss curves to analyze model performance.

Task 2: AI Agent with RAG and Function Calling

Important : I couldn't run the Hugging Face API because it is taking time to take my submission approval for Meta LLama 2 API integration access

Objective

Develop an AI agent capable of processing files as input using LlamaIndex or LangChain. The agent should exhibit reasoning capabilities, support Retrieval-Augmented Generation (RAG), and implement function calling for arithmetic operations.

Features Implemented

• File Ingestion using LlamaIndex or LangChain.
• Retrieval-Augmented Generation (RAG) to enhance responses with retrieved information.
• Function Calling:
  • Arithmetic operations: Addition, Subtraction, Multiplication, Division.
• Reasoning Capabilities:
  • The agent processes and infers information from provided files.

Usage

CIFAR-10 Classification

Run the notebook GDSC_Task_1.ipynb to train and evaluate the CNN model.

AI Agent with RAG

Run the notebook GDSC_Task_2.ipynb to interact with the AI agent.