🧠 Deep Learning Applications

A comprehensive, hands-on collection of 90+ deep learning programs — from single neurons to Transformers, CNNs, RNNs, AI Agents, and real-world applications.

Explore Programs · Getting Started · Real-World Projects

---

📌 Overview

This repository is a progressive deep learning curriculum implemented entirely in Python. It covers the full journey from bare-metal neuron math to production-ready architectures — every concept is implemented step-by-step with graphical and animated visualizations wherever possible.

The collection spans 7 major tracks:

Track	Files	Topics
🔵 ANN	1–12	Single neurons → multi-layer training → animated backprop
🟢 RNN	1–21, 43–70	Tokenization → embeddings → LSTM → sentiment analysis
🔴 CNN	22–29, 39–42, 46	Pixel ops → edge detection → crack detection
🟡 Loss Functions	13–17	MSE, MAE, Binary Cross-Entropy
⚡ TensorFlow/Keras	30–38	Tensor ops → neural nets → image classification
🟣 Transformers	71–78	Self-attention → encoder-decoder → GPT-style decoder
🤖 AI Agents & Apps	79–90	Streamlit UIs → AI agents → LLM-powered Q&A

---

🚀 Getting Started

Prerequisites

python >= 3.8

Installation

# Clone the repository
git clone https://github.com/Jayesh-011/Deep_Learning_Applications.git
cd Deep_Learning_Applications

# Install dependencies
pip install tensorflow keras numpy pandas matplotlib scikit-learn streamlit

Run any program

python 1_ANN_Single_Neuron.py
python 48_Reliance_LSTM_Time_Series.py
streamlit run 86__Intelligent_Document_Question_Answering_System.py

---

📂 Program Index

🔵 ANN — Artificial Neural Networks

File	Description
1_ANN_Single_Neuron.py	Manual single neuron with weights and bias
2_Activation_Relu.py	ReLU activation — manual + graphical
3_Activation_Sigmoid.py	Sigmoid activation — manual + graphical
4_Activation_Sigmpoid_vs_Relu.py	Side-by-side comparison plot
5_Layered_Artificial_Neural_Network.py	Full layered ANN from scratch
6_Layered_Artificial_Neural_Network_Simplified.py	Cleaner modular version
7_Layered_Artificial_Neural_Network_Graphical.py	Network visualization
8_ANN_Trianing_Steps.py	Forward pass + manual backprop
9_ANN_Trianing_Steps_Graphical.py	Training steps with live plots
10_ANN_MultiNeuron_Trianing_Steps.py	Multi-neuron training walkthrough
10_ANN_MultiNeuron_Trianing_Steps_Multi_Valuee.py	Batch training extension
11_ANN_MultiNeuron_Trianing_Steps_Graphical.py	Graphical multi-neuron training
12_ANN_MultiNeuron_Trianing_Steps_Graphical_Animation.py	Animated weight update visualization
16_Gradient_Backprapogation.py	Gradient descent & backprop manual
17_Gradient_Backprapogation_Graphical.py	Graphical gradient flow

🟢 RNN — Recurrent Neural Networks

File	Description
1_RNN_DatasetDisplay.py	Dataset loading and display
2_RNN_Tokenization.py	Word tokenization from scratch
3_RNN_VocabularyCreation.py	Vocabulary builder
4_RNN_KerasTokenization.py	Tokenization using Keras
5_RNN_TexttoSequences.py	Text → integer sequences
6_RNN_UseOfPadding.py	Padding sequences manually
7_RNN_PaddingKeras.py	Keras pad_sequences
8_RNN_VocabularySize.py	Vocabulary size management
9_RNN_OutputLabels.py	Label encoding for RNN output
10_RNN_EmbeddingManual.py	Manual word embedding vectors
11_RNN_EmbeddingKeras.py	Keras Embedding layer
12_RNN_TimeSteps.py	Understanding time steps
13_RNN_HiddnState.py	Hidden state mechanics
14_RNN_ManualCalculations.py	Full RNN forward pass by hand
15_RNN_Tanh.py	Tanh activation in RNNs
16_RNN_Sigmoid.py	Sigmoid gating demo
17_RNN_BinaryCrossEntropyLoass.py	BCE loss for sequence models
18_RNN_Architecture.py	Keras RNN architecture build
19_RNN_TrainModel.py	Training loop
20_RNN_PredictionFlow.py	End-to-end prediction pipeline
21_RNN_CompleteCode.py	Full RNN from scratch to prediction
43_RNN_Character_Prediction.py	Character-level text generation
44_RNN_Sentiment_Analysis.py	Movie review sentiment classifier
45_RNN_Sequance_Prediction.py	Numeric sequence forecasting
70_RNN_Graphical_Version.py	Animated RNN architecture visualizer

Files 49–69 mirror files 1–21 as extended/revised versions.

🔴 CNN — Convolutional Neural Networks

File	Description
22_Convert28by28_Grayscale_Image.py	Image resizing to 28×28 grayscale
23_Pixel_Grayscale_Display.py	Pixel matrix display
24_Color_Image_Pixel_Display.py	RGB channel visualization
25_CNN_Edge_Detection.py	Manual convolution edge filter
26_CNN_Edge_Detection_Result_Display.py	Edge detection output display
27_CNN_Edge_Detection_Result_Animation.py	Animated kernel sliding
28_CNN_Edge_Detection_Cat.py	Edge detection on real image
29__Realtime_Image_Classification.py	Real-time image classifier
39_CNN_Convolution_ReLU_Pooling_FC.py	Full CNN pipeline
40_CNN_Convolution_ReLU_Pooling_FC_Graphical.py	Graphical CNN layer-by-layer
41_CNN_Relu_Pooling_flattern.py	Feature map flattening
42_CNN_Convolution_ReLU_Pooling_FC_Graphical_BigData.py	CNN on large input
46__CNN_Surface_Crack_Detection.py	Industrial surface crack classifier

🟡 Loss Functions

File	Description
13_LossFunction_MSE.py	Mean Squared Error
14_LossFunction_MAE.py	Mean Absolute Error
15_LossFunction_Binary_Cross_Entropy.py	Binary Cross-Entropy

⚡ TensorFlow / Keras Fundamentals

File	Description
30_tensorflow_tensor_types.py	Tensor types: constant, variable, sparse
31_tensorflow_tensor_operations.py	Add, subtract, multiply ops
32_tensorflow_tensor_shape_reshape.py	Shape inspection and reshape
33_tensorflow_tensor_variable.py	Mutable tf.Variable
34_tensorflow_tensor_matrix_multiplication.py	Matrix multiplication with tf
35_tensorflow_tensor_single_neuron.py	Single neuron using TensorFlow
36_tensorflow_tensor_neural_network.py	MLP with TensorFlow
37_tensorflow_tensor_calassification.py	Binary classification model
38_tensorflow_tensor_ANN_image_calassification.py	ANN for image classification

🟣 Transformers

File	Description
71_transformer_self_attention.py	Scaled dot-product self-attention
72_transformer_Multihead_attention.py	Multi-head attention mechanism
73_transformer_poitionalencoder.py	Sinusoidal positional encoding
74_transformer_encoder.py	Full transformer encoder block
75_transformer_decoder.py	Full transformer decoder block
76_Transformer_Sentiment_Classification_Encoder_Only.py	Encoder-only sentiment model
77_Transformer_Neural_Machine_Translation_Encoder_Decoder.py	Seq2seq NMT model
78_Transformer_Generative_Pre_trained_Transformer_Decoder_Only.py	GPT-style decoder

🤖 AI Agents & Applications

File	Description
79_streamlit_JayGanesh.py	Hello World Streamlit app
80_streamlit_Input.py	Streamlit text input widget
81_streamlit_Button.py	Streamlit button interaction
82_streamlit_fileUpload.py	File upload handler
83_streamlit_TextExtraction.py	Extract text from uploaded docs
84_streamlit_Chunking.py	Text chunking for LLM context
85__Calculator_tkinter.py	GUI Calculator with tkinter
86__Intelligent_Document_Question_Answering_System.py	LLM-powered PDF Q&A system
87_AIAgent_RuleBased.py	Rule-based AI agent
88_AIAGentCalculator.py	AI agent with calculator tool
89_AIAGenyMultpleTool.py	Multi-tool AI agent
90_AIAgent_Memory.py	AI agent with persistent memory
interview_agent.py	AI interview simulation agent

---

🏆 Real-World Projects

📈 Reliance Stock Price Prediction (LSTM)

File: 48_Reliance_LSTM_Time_Series.py
Time-series forecasting using stacked LSTM on Reliance Industries stock data. Outputs prediction charts and CSV results.

Assets: _reliance_stock_sample.csv · _actual_vs_predicted_detailed.png · _training_loss_detailed.png

---

🏥 Breast Cancer Classification (FNN)

File: 47_IndustrialBreastCancer.py
Feedforward neural network on the Wisconsin Breast Cancer dataset for malignant/benign classification.

Asset: breast-cancer-wisconsin.csv

---

🏗️ Surface Crack Detection (CNN)

File: 46__CNN_Surface_Crack_Detection.py
Convolutional neural network to classify concrete surface images as cracked or intact using a custom image dataset.

Asset: CrackDataset/

---

🎓 Student Placement Prediction (FNN)

Files: 19_FNN_Student_Result_Classification.py, 21_FNN_Placement_Prediction.py
Classification models predicting student pass/fail and campus placement probability.

Asset: student.csv · placement_data.csv

---

📄 Intelligent Document Q&A System

File: 86__Intelligent_Document_Question_Answering_System.py
A Streamlit-powered RAG pipeline that lets users upload documents and ask questions answered by an LLM.

---

🛠️ Tech Stack

Library	Purpose
TensorFlow / Keras	Neural network building and training
NumPy	Matrix math, manual implementations
Pandas	Data loading and preprocessing
Matplotlib	Plots, animations, visualizations
Scikit-learn	Preprocessing, metrics
Streamlit	Interactive web applications
tkinter	Desktop GUI

---

📁 Repository Structure

Deep_Learning_Applications/
├── 1–12    ─── ANN (Artificial Neural Networks)
├── 13–17   ─── Loss Functions
├── 1–21    ─── RNN Series I (NLP Foundations)
├── 22–29   ─── Image Processing & CNN Basics
├── 30–38   ─── TensorFlow Fundamentals
├── 39–42   ─── CNN Advanced
├── 43–45   ─── RNN Applications
├── 46–48   ─── Industrial / Real-World Projects
├── 49–70   ─── RNN Series II (Extended)
├── 71–78   ─── Transformers
├── 79–90   ─── AI Agents & Streamlit Apps
├── CrackDataset/           ← Surface crack images
├── breast-cancer-wisconsin.csv
├── placement_data.csv
├── student.csv
└── _reliance_stock_sample.csv

---

🤝 Contributing

Contributions are welcome! To add a new program:

1. Fork the repository
2. Follow the existing naming convention: NN_Topic_Description.py
3. Add a brief docstring at the top of your file
4. Submit a pull request

---

👨‍💻 Author

Jayesh — @Jayesh-011

---

⭐ Star this repo if it helped you learn deep learning!