Projects
Image Caption Generation [GitHub]
Utilized the VGG16 model to extract high-level features from images.
Implemented an encoder-decoder framework with CNN for encoding and LSTM for caption generation.
Tokenized and preprocessed caption data including sequence padding and embeddings.
Trained the model with a custom data generator using categorical crossentropy loss and Adam optimizer.
Implemented an encoder-decoder framework with CNN for encoding and LSTM for caption generation.
Tokenized and preprocessed caption data including sequence padding and embeddings.
Trained the model with a custom data generator using categorical crossentropy loss and Adam optimizer.
Language Translation [GitHub]
Designed an encoder-decoder architecture with LSTM layers for sequence-to-sequence translation.
Encoder generates a fixed-size context vector; decoder generates translated output.
Used embeddings and LSTM for capturing long-term dependencies.
Encoder generates a fixed-size context vector; decoder generates translated output.
Used embeddings and LSTM for capturing long-term dependencies.
Auto Complete [GitHub]
Built using an n-gram language model to predict the next word based on context.
Computes perplexity scores of candidate words and selects the most probable one.
Provides real-time word suggestions to improve typing efficiency.
Computes perplexity scores of candidate words and selects the most probable one.
Provides real-time word suggestions to improve typing efficiency.
PDF Summarizer [GitHub]
Extracts text directly from PDFs or uses OCR for scanned files.
Performs word frequency analysis and sentence scoring for summarization.
Applies stemming, tokenization, and stopword removal.
Includes a command-line interface for flexible usage.
Performs word frequency analysis and sentence scoring for summarization.
Applies stemming, tokenization, and stopword removal.
Includes a command-line interface for flexible usage.