Independent publisher
Jeremy
jermo1120
Compare the published evidence
This profile connects the 4 model and dataset records shown below to their upstream cards and sources. The library counts describe SAVRN's coverage, rather than this publisher's entire output.
Published tasks: Robotics.
Reported licenses: apache-2.0. Open a record to review its specific license and conditions.
Compare model specifications, source dates, and hardware estimates on the individual records. Publisher identity verification is separate from a model's quality, deployment readiness, or suitability for your work.
Models
This is a single-task vision-language-action policy for task 38 sprayingforbugs of the BEHAVIOR-1K benchmark. It is Pi0.5-based, has 3.4B parameters, and controls the R1Pro robot from three 224×224 RGB cameras plus proprioception. It predicts 30-step action chunks. It was fine-tuned for 25,000 steps on task 38 demonstrations only. The starting point was checkpoint3 of the 1st-place 2025 BEHAVIOR Challenge solution (Larchenko, Zarin, Karnatak; paper). checkpoint3 had been trained on 13 tasks (4, 27, 31–33, 35–39, 41, 46, 49), task 38 among them. Status: training only. This model has not been evaluated in the BEHAVIOR simulator yet. Lower training loss does not guarantee a higher task success…
Datasets
Rollouts of the Pi0.5 fine-tune jermo1120/b1k-task55-re-shelving-library-books (step24999) on BEHAVIOR-1K task reshelvinglibrarybooks, following the 2026 challenge leaderboard protocol. Missing instances count as 0 in Q. qscore gives partial credit: 1.0 for a full success, otherwise the fraction of goal predicates that were false at the start and true at the end of the episode. The task 55 goal is only "every book (3) is inside a bookcase", so per-instance scores are 0, 0.33, 0.67 or 1.0. Distances are normalized by the mean human demo (1.0 = same path length as a human). Per-rollout metrics are in results/ /.json and results/summary.csv, videos (head + both wrist cameras) in videos/ /…