Dhara Foundational Models - a codelion Collection

codelion 's Collections

Dhara Foundational Models

Sutra Pedagogical Datasets

Nano Language Models

Pre-training Dataset Samples

Pivotal Token Search

Internal Coherence Maximization

Dhara Foundational Models

updated May 26

Diffusion Language Models combining deep narrow networks, Canon layers (depthwise causal convolutions), and WSD (Warmup-Stable-Decay) training.

Running

Agents

1

dhara-chat

🌊

1

CPU demo of dhara-250M tri-mode (AR/diffusion/self-spec)
codelion/dhara-250m

Text Generation • 0.2B • Updated 15 days ago • 244 • 3
codelion/dhara-250m-ar-base

Text Generation • 0.2B • Updated 15 days ago • 18 • 1
codelion/dhara-70m

Text Generation • 71.3M • Updated Dec 30, 2025 • 330 • 49