YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
LBS on DMax: routers
EB-Decoding NewRouter checkpoints for DMax-16B (state dict without the frozen token embedding; load with train_router_sdar.load_router(path, 32, device)).
ckpt_inblock_op/router.pt: DMax-16B trajectories, CE<1.0 (GSM8K headline: 91.13 / 5.11 TPF).ckpt_op_ce0.3/router.pt: DMax-16B trajectories, CE<0.3 (selected on MBPP).ckpt_inblock_mc,ckpt_mc_ce*: LLaDA-2.0-mini trajectories, math+code.ckpt_inblock: math only.ckpt_sdar: SDAR-path port (W=32/64).ckpt: early rebuilt router.
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐ Ask for provider support