Nabidnur commited on
Commit
654707f
·
verified ·
1 Parent(s): c7608b2

RoboRACE: code, cached agent outputs, results, media, SO-101 renders (part 4)

Browse files
scripts/make_readme.py CHANGED
@@ -108,6 +108,8 @@ def main():
108
 
109
  [Leaderboard](results/LEADERBOARD.md) · [Protocol](docs/BENCHMARK.md) · [Plan](docs/PLAN.md) · [Architecture](#architecture) · [Videos](media/videos) · [Figures](media/figures)
110
 
 
 
111
  <img src="media/videos/robot_rollouts.gif" width="960" alt="SO-101 arm rollouts seen by five camera agents; three cameras lie; majority vote flips, RACE-S stays correct">
112
 
113
  <sub>A MuJoCo SO-101 arm (Menagerie model) runs pick-and-place; ground truth is the final physics state. Five camera agents judge success with
@@ -307,7 +309,8 @@ src/robotics/benchmarks/ erqa.py · blocksworld.py (exact simulator + validator
307
  src/robotics/bench/ trackA/B/C · erqa_swarm · planbench_eval · leaderboard (bootstrap ranks)
308
  src/robotics/sim/ so101_env.py (SO-101 IK pick-and-place, physics ground truth) · tabletop.py
309
  src/robotics/vlm/ cosmos_reason.py (logit P(yes) verifier)
310
- scripts/ run_benchmark · make_figures · make_videos · make_gifs · make_architecture · make_plotnn · make_readme · agents
 
311
  results/ LEADERBOARD.md · leaderboard.json · cached agent outputs
312
  media/ figures · videos (mp4 + gif) · architecture (Raschka-style + PlotNeuralNet)
313
  ```
 
108
 
109
  [Leaderboard](results/LEADERBOARD.md) · [Protocol](docs/BENCHMARK.md) · [Plan](docs/PLAN.md) · [Architecture](#architecture) · [Videos](media/videos) · [Figures](media/figures)
110
 
111
+ [🤗 Dataset · full work](https://huggingface.co/datasets/Nabidnur/RoboRACE-Bench) · [🤗 Space · showcase](https://huggingface.co/spaces/Nabidnur/RoboRACE)
112
+
113
  <img src="media/videos/robot_rollouts.gif" width="960" alt="SO-101 arm rollouts seen by five camera agents; three cameras lie; majority vote flips, RACE-S stays correct">
114
 
115
  <sub>A MuJoCo SO-101 arm (Menagerie model) runs pick-and-place; ground truth is the final physics state. Five camera agents judge success with
 
309
  src/robotics/bench/ trackA/B/C · erqa_swarm · planbench_eval · leaderboard (bootstrap ranks)
310
  src/robotics/sim/ so101_env.py (SO-101 IK pick-and-place, physics ground truth) · tabletop.py
311
  src/robotics/vlm/ cosmos_reason.py (logit P(yes) verifier)
312
+ scripts/ run_benchmark · make_figures · make_videos · make_gifs · make_architecture · make_plotnn · make_readme ·
313
+ publish_hf + make_space (Hugging Face dataset card + Space) · agents
314
  results/ LEADERBOARD.md · leaderboard.json · cached agent outputs
315
  media/ figures · videos (mp4 + gif) · architecture (Raschka-style + PlotNeuralNet)
316
  ```
scripts/make_space.py CHANGED
@@ -20,7 +20,7 @@ def pct(x, d=1):
20
 
21
 
22
  def table(rows, cols, highlight=lambda r: False):
23
- head = "".join(f"<th>{html.escape(c)}</th>" for c, _ in cols)
24
  body = ""
25
  for r in rows:
26
  cls = ' class="ours"' if highlight(r) else ""
 
20
 
21
 
22
  def table(rows, cols, highlight=lambda r: False):
23
+ head = "".join(f"<th>{c}</th>" for c, _ in cols) # column titles are trusted literals (may contain entities)
24
  body = ""
25
  for r in rows:
26
  cls = ' class="ours"' if highlight(r) else ""
scripts/publish_hf.py CHANGED
@@ -213,6 +213,7 @@ def stage():
213
 
214
 
215
  def publish_space(api, token):
 
216
  from make_space import build
217
  sp = build(STAGE.parent / (STAGE.name + "_space"))
218
  api.create_repo(SPACE, repo_type="space", space_sdk="static", exist_ok=True, token=token)
@@ -224,7 +225,13 @@ def main():
224
  ap = argparse.ArgumentParser()
225
  ap.add_argument("--dry-run", action="store_true")
226
  ap.add_argument("--skip-space", action="store_true")
 
227
  a = ap.parse_args()
 
 
 
 
 
228
  stage()
229
  if a.dry_run:
230
  return
 
213
 
214
 
215
  def publish_space(api, token):
216
+ sys.path.insert(0, str(Path(__file__).resolve().parent))
217
  from make_space import build
218
  sp = build(STAGE.parent / (STAGE.name + "_space"))
219
  api.create_repo(SPACE, repo_type="space", space_sdk="static", exist_ok=True, token=token)
 
225
  ap = argparse.ArgumentParser()
226
  ap.add_argument("--dry-run", action="store_true")
227
  ap.add_argument("--skip-space", action="store_true")
228
+ ap.add_argument("--space-only", action="store_true")
229
  a = ap.parse_args()
230
+ if a.space_only:
231
+ from huggingface_hub import HfApi
232
+ publish_space(HfApi(), os.environ["HF_TOKEN"])
233
+ print("space:", f"https://huggingface.co/spaces/{SPACE}")
234
+ return
235
  stage()
236
  if a.dry_run:
237
  return