divelab/opdlm_eval_data
Viewer • Updated • 8.52k • 290
OPDLM
Data and checkpoints for Data-Efficient Autoregressive-to-Diffusion Language Models via On-Policy Distillation
-
-
divelab/opdlm_train_data
Viewer • Updated • 61.8k • 147 -
divelab/Qwen3-4B-a2d-init
Text Generation • 4B • Updated • 35 • 1 -
divelab/Qwen3-8B-a2d-init
Text Generation • 8B • Updated • 31 -
divelab/Qwen3-1.7B-a2d-init
2B • Updated • 43 -
divelab/Qwen3-0.6B-a2d-init
0.6B • Updated • 44 -
divelab/OPDLM-8B
Text Generation • 8B • Updated • 181 • 1 -
divelab/OPDLM-4B
Text Generation • 4B • Updated • 9.85k -
divelab/OPDLM-1.7B
Text Generation • 2B • Updated • 23 -
divelab/OPDLM-0.6B
Text Generation • 0.6B • Updated • 11.5k -
Data-Efficient Autoregressive-to-Diffusion Language Models via On-Policy Distillation
Paper • 2606.06712 • Published • 2 -
divelab/OPDLM-MATH-4B
Text Generation • 4B • Updated • 2.98k -
divelab/OPDLM-MATH-8B
Text Generation • 8B • Updated • 52 -
divelab/OPDLM-MATH-4B-Thinking
Text Generation • 4B • Updated • 48 -
divelab/OPDLM-MATH-8B-Thinking
Text Generation • 8B • Updated • 59