AI & LLMs

Mini Language Model with TensorFlow

Transformer-based language model trained from scratch on the Tiny Shakespeare dataset for next-token prediction using custom tokenization and multi-head self-attention.

Key Architectural & System Features

Designed transformer model with custom tokenization and multi-head self-attention
Trained using next-token prediction objective
Generated coherent text continuations via greedy and temperature-based sampling

Technologies & Tools Stack

TensorFlowKerasTransformersTiny Shakespeare Dataset