# seohongpark/ocbench

> OCBench is a controllable benchmark for robotic manipulation, designed for studying behavioral cloning (BC) and reinforcement learning (RL) algorithms. It provides GPU-accelerated tasks and scripted policies that mimic human demonstrations, enabling controlled studies with virtually unlimited data.

- Magnitude: 4.1 out of 10 — Early signal
- Stars: 62 total · +2 stars measured 2026-10-07, 02:25–04:29 UTC, ≈ 36 by evening
- Star trust: star growth looks organic
- Category: Science & research · Language: Python · License: MIT · Created: 2026-10-05 · Last push: 2026-10-07
- GitHub: https://github.com/seohongpark/ocbench · Homepage: https://seohong.me/projects/ocbench/ · Page: https://gitnova.dev/en/r/seohongpark/ocbench

## Useful for

- Run 1024 parallel environments to collect reinforcement learning data
- Train and compare BC and RL algorithms on 28 manipulation tasks
- Generate demonstration datasets using scripted policies

## Why it’s here

- Star-counter measurements on 2026-10-07 (UTC), 02:25–04:29: 60 → 62 stars (+2). This is the change over that interval.
- Estimated end-of-day forecast: about +36 stars, using observed gains and the previous day.
- The repository is 2 days old and already has 62 stars. With less than two weeks of history, there's no usual pace to compare the spike against yet.
- Top new repositories this week: #183.
- About 6 forks a day — people are taking the code.

## Star trust

Star growth looks organic.

Star-trust labels are heuristics based on the repository’s behavior, not a check of every stargazer.

## Numbers

- Forks: 6
- Issues and pull requests: 2
- Watchers: 1
- Average over the last week: 33 per day
- Usual pace: too little history (under two weeks)
- Stars in the last hour (measured): 1
- Latest release: v1.0.1 (2026-10-06)

## Stars per day, last 4 days (oldest → newest, today is partial)

2026-10-04 … 2026-10-07: 0, 12, 50, 2

## Spotted in now

- Top new repositories this week: #183

## Similar by description

1. **BytedTsinghua-SIA/DAPO** — 0.4 · Quiet · Language models · Python · +0 stars measured 2026-10-07, 00:29–04:38 UTC, ≈ 1 by evening
   Open-source reinforcement learning system for large-scale LLM training, including the DAPO algorithm, code infrastructure, and dataset, plus released DAPO-Qwen-32B weights.
   Full card: https://gitnova.dev/en/r/BytedTsinghua-SIA/DAPO.md
2. **apromisedland/VLCoT** — 0.0 · Steady · Science & research · Python · Not enough measurements today
   Research implementation of VLCoT: stage-verifiable latent chain-of-thought reasoning with observation-grounded verification and dependency-aware local repair for long-horizon robotic manipulation. Includes an OpenVLA-7B adapter, LoRA…
   Full card: https://gitnova.dev/en/r/apromisedland/VLCoT.md
3. **zjwzcx/Awesome-Astra-Embodied-AI** — 1.1 · Steady · Learning & lists · +2 stars measured 2026-10-07, 00:27–04:36 UTC, ≈ 8 by evening
   A curated collection of demos and cases of GPT-6 Astra in embodied AI and robotics: zero-shot control, deployment in simulation and on real robots, RL training, and real-to-sim replay.
   Full card: https://gitnova.dev/en/r/zjwzcx/Awesome-Astra-Embodied-AI.md

---

Magnitude (0–10) measures how fast and how unusually interest in a repository is growing right now. It is not a quality score.
Days are UTC. “So far today” is a fact; “expected by the end of the day” is a forecast.
Summaries and use cases are written by an LLM (DeepSeek V4.1 Flash) from the README and may be inaccurate: verify specific claims (benchmarks, speed, hardware) in the repository itself.
Data as of 2026-10-07 04:39 UTC, updated every 30 minutes.