pufanyi/vbvrpro_sampler_trajectories-data

VBVR-Pro sampler trajectory media

This media archive backs the interactive `pufanyi/vbvrpro_sampler_trajectories` Space.

It contains 12 matched evaluation cells:

• —DiffSynth step-35500 baseline and DanceGRPO checkpoint 2200
• —Flow-CPS noise 0.1, 0.3, 0.7, and 0.9
• —deterministic FlowMatch Euler ODE and UniPC ODE
• —500 samples per cell across 100 VBVR-Pro tasks

The deployment is split across three public media repositories so each Git-backed Dataset remains below Hugging Face's recommended 100,000-file threshold:

• —`pufanyi/vbvrpro_sampler_trajectories-data`: optional 6×5 overview grids and native final outputs
• —`pufanyi/vbvrpro_sampler_trajectories-baseline-steps`: 90,000 original baseline step videos
• —`pufanyi/vbvrpro_sampler_trajectories-2200-steps`: 90,000 original checkpoint-2200 step videos

For each of the 6,000 model/sampler/sample combinations, the complete deployment stores all 30 native step_00.mp4 through step_29.mp4 files without spatial repacking, plus:

• —steps_grid.mp4: a compressed synchronized 6×5 overview
• —final_00.mp4: the native 512×512, 81-frame final output at sigma zero

Paths follow:

videos/{cell_id}/{domain_folder}/{task_name}/{sample_id}/{filename.mp4}

The companion data/index.json records sample identity, exact prompts, seeds, sampler schedules, trajectory semantics, and 6,000 aligned final-only EvalKit scores on a 0–1 scale. A displayed score evaluates public step 30 / final_00.mp4, not the intermediate clean-endpoint previews. No training data or model weights are included.