This repository contains the latest 30,000-step post-training checkpoint from the StarVLA MiniCPM-GR00T experiment on the BananaInBowl HDF5 dataset. The main weight file is steps30000pytorchmodel.pt. The model was initialized from the 145,000-step DROID pretraining checkpoint and trained with two RGB views resized directly to 448x448, a 15-step action horizon, and 8-dimensional state/action vectors. This is a StarVLA training checkpoint rather than a standalone Transformers frompretrained model. Use the matching StarVLA code and the included config.full.yaml to load it.
Independent publisher
Tingrui Shen
trshen925
Models in Library1
Datasets in Library0
Models on Hugging Face6
Followers—
Compare the published evidence
This profile connects the 1 model and dataset record 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.