docs: update card for AssembleBench rename

README.md

@@ -15,14 +15,18 @@ configs:
   data_files: data/*/*.parquet
 ---
 
-# assembly_bench_3
+# AssembleBench
 
-Multi-task contact-rich assembly demonstrations for imitation learning / VLA fine-tuning.
-Built from a cleaned and topped-up revision of [`assembly_bench_2`](https://huggingface.co/datasets/lukasskellijs/assembly_bench_2), with a second QC pass for duplicates, idle chunks, length outliers, and within-family Mahalanobis outliers.
+Multi-task contact-rich assembly demonstrations for imitation learning / VLA fine-tuning,
+and the training set behind the [AssembleBench](https://github.com/hud-evals/assemble-bench)
+benchmark: 14 tasks on the NIST ATB-1 taskboard (insert round and square pegs, mesh gears,
+thread nuts) on the DROID robot platform.
+
+Read the writeup: [Benchmarking Robot Models on Contact-Rich Assembly](https://www.hud.ai/blog/assemble-benchmark).
 
 This dataset was created using [LeRobot](https://github.com/huggingface/lerobot).
 
-<a class="flex" href="https://huggingface.co/spaces/lerobot/visualize_dataset?path=lukasskellijs/assembly_bench_3">
+<a class="flex" href="https://huggingface.co/spaces/lerobot/visualize_dataset?path=hud-evals/AssembleBench">
 <img class="block dark:hidden" src="https://huggingface.co/datasets/huggingface/badges/resolve/main/visualize-this-dataset-xl.svg"/>
 <img class="hidden dark:block" src="https://huggingface.co/datasets/huggingface/badges/resolve/main/visualize-this-dataset-dark-xl.svg"/>
 </a>
@@ -54,22 +58,25 @@ Language-conditioned pick / insert / mesh / thread:
 
 ## How it was built
 
-Pipeline:
+Built from a cleaned and topped-up revision of
+[`lukasskellijs/assembly_bench_2`](https://huggingface.co/datasets/lukasskellijs/assembly_bench_2),
+with a second QC pass for duplicates, idle chunks, length outliers, and within-family
+Mahalanobis outliers. Pipeline:
 
-1. Start from [`assembly_bench_2_clean`](https://huggingface.co/datasets/lukasskellijs/assembly_bench_2_clean) (911 eps: dups + long outliers already removed from the original 1500).
+1. Start from the cleaned v2 revision (911 eps: dups + long outliers already removed from the original 1500).
 2. Drop all **M20** nut episodes.
 3. Append scripted-expert top-up HDF5 demos that refill the gaps left by the clean pass (success-filtered Isaac Lab recordings).
 4. QC blocklist (45 trajectories by MD5 of full `action || state`):
    - exact-duplicate extras (none remained after clean + ingest dedup)
-   - idle runs ≥ 15 frames (action-chunk size used in training)
-   - per-task length >3σ and clear within-family **20-PC Mahalanobis** outliers (retry / thrash / bad seat)
+   - idle runs >= 15 frames (action-chunk size used in training)
+   - per-task length >3 sigma and clear within-family **20-PC Mahalanobis** outliers (retry / thrash / bad seat)
 5. Final exact-duplicate scan must be zero or the build aborts.
 
 Quality over quantity: mild multivariate peg shape outliers that look like valid successes were kept.
 
 ## Schema
 
-Same contract as `assembly_bench_2` / DROID joint-pos:
+DROID joint-position contract:
 
 | key | shape | notes |
 |-----|-------|--------|
@@ -86,7 +93,7 @@ Same contract as `assembly_bench_2` / DROID joint-pos:
 ```python
 from lerobot.datasets.lerobot_dataset import LeRobotDataset
 
-ds = LeRobotDataset("lukasskellijs/assembly_bench_3")
+ds = LeRobotDataset("hud-evals/AssembleBench")
 print(ds.num_episodes, ds.num_frames)
 print(ds[0]["observation.images.wrist"].shape)
 ```
@@ -98,8 +105,16 @@ print(ds[0]["observation.images.wrist"].shape)
 
 Not a teleop human dataset. Demonstrations are from a privileged scripted expert in Isaac Sim / Isaac Lab; visuals and proprio match the policy observation contract.
 
-## Citation / lineage
+## Reference checkpoints
+
+pi0.5 finetunes trained on this dataset:
+
+| Checkpoint | What it is |
+|---|---|
+| [`pi05-AssembleBench-12k`](https://huggingface.co/hud-evals/pi05-AssembleBench-12k) | behavior-cloning baseline, all 14 tasks |
+| [`pi05-AssembleBench-cgdagger-r3`](https://huggingface.co/hud-evals/pi05-AssembleBench-cgdagger-r3) | BC + 3 rounds of code-gated DAgger, the final checkpoint |
+
+## Lineage
 
-- Source family: `lukasskellijs/assembly_bench_2` → `assembly_bench_2_clean` → top-up + QC → **`assembly_bench_3`**
 - Simulator: NVIDIA Isaac Sim 6 / Isaac Lab Arena
 - Format: LeRobot v3.0