lyricsLM a transformer, running on a tiny server

Type a line. Watch it become verse.

Generated by a 2.7M parameter transformer, trained from scratch on public-domain poetry, Tagore, Naidu, Kabir, Wordworth, Tennyson,Blake, Poe, and more. The model runs on a small cpu server.

Give it a line of poetry, or just an idea, and it'll continue in its own voice. Try "and miles to go before I sleep".
Creativity 1.1 · loose
Ready.

What's actually happening

01

Runs on a small server

The model runs server-side — a few million parameters, milliseconds per request. No download, no setup, always ready.

02

Architecture

Embeddings, attention, softmax, sampling. See my article and code using the link below. The model architecture follows Andrej Karpathy's "Let's Build GPT", a well established way to build a transformer.

03

Public-domain training data

The poetry it learned from are Tagore, Naidu, Kabir, Dickinson, Poe, and more. The data comes from Project Gutenberg. Every book is in the public domain, royalty free, and free of copyright.