Skip to content

adithyan-sys/Training-a-LLM-model-

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Training-a-LLM-model

TinyGPT - A Simple GPT-style Language Model from Scratch

A minimal implementation of a decoder-only Transformer (GPT-like) language model trained on a small custom corpus. Built for learning purposes using PyTorch.

Perfect for beginners who want to understand how GPT models work under the hood.

Features

  • From Scratch Implementation: No high-level libraries like Hugging Face
  • Multi-Head Self-Attention with causal masking
  • Pre-Norm Transformer Blocks
  • Positional Embeddings
  • Train/Validation Split with proper evaluation
  • Top-k Sampling + Temperature for better text generation
  • Gradient Clipping and Dropout for stability
  • Colab Compatible (single notebook version available)
  • Model saving and easy inference

About

In this project i have leant to train a GPT - LLM model from open resources ,in this project we give a raw data to train the model

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors