Diffusion Language Models combining deep narrow networks, Canon layers (depthwise causal convolutions), and WSD (Warmup-Stable-Decay) training.
-
dhara-chat
🌊1ZeroGPU demo of dhara-250M tri-mode (AR/diffusion/self-spec)
-
codelion/dhara-250m
Text Generation • 0.2B • Updated • 2.02k • 4 -
codelion/dhara-250m-ar-base
Text Generation • 0.2B • Updated • 47 • 1 -
codelion/dhara-70m
Text Generation • 71.3M • Updated • 192 • 48