Kite Trossen Sim Cup All ACT 30k

Offizielle Gewichte
Fachartikel und technische Berichte

Kite's task-trained LeRobot ACT policy for cup picking, trained on seven simulated datasets for 30,000 steps. This records the author's revised simulation evaluation, with unchanged weights.

Vollständige Modellbeschreibung ansehen

Fachartikel und technische Berichte

Fachartikel / Bericht unbestätigt

Offizielle Informationen sind erfasst; ein versionsspezifischer Fachartikel oder Bericht wurde noch nicht bestätigt.

Zuletzt geprüft: 2026-11-18

Zugang und Nutzung

Öffentliche Code-Ressourcen
Öffentlicher Zugang
Modelldateien
Öffentlicher Zugang
Trainingsdaten
Öffentlicher Zugang

EvaluationsunterlagenEvaluationsunterlagen

Trainings- / Fine-Tuning-Code

Zugang
Öffentlicher Zugang

QuelleGeprüft am · 2026-11-18

Code-Lizenz
Apache-2.0

Modelldateien

Modelldateien
trossen-bin-picking-sim-act-cup-all · 30000ZugangGewichtelizenz
Veröffentlichungsangaben

QuelleGeprüft am · 2026-11-18

Vor der Bereitstellung prüfen

Author simulation only; two layouts; changed protocol and render-domain bias; no hardware or independent reproduction. Verfügbare Gewichte oder Code belegen keine sofortige Roboterkompatibilität. Prüfen Sie Konfiguration, Daten und Schnittstellen in der Quelldokumentation.

Unterlagen zu Bereitstellung und Nutzung

Inferenzmodus
ACTPolicy.from_pretrained; LeRobot >=0.6; pin commit
Laufzeit und Frameworks
lerobot-pytorch
Bereitstellungsanforderungen
Four RGB cameras and 14 joint positions on bimanual WidowX AI

Architektur und Ein-/Ausgaben

Modelltyp
Aktionspolicy
Architektur
Transformer
Backbone
ResNet18 (ImageNet1K V1), Transformer 512, encoder 4 / decoder 1
Eingabe
RGB-Bilder / Roboterzustand
Ausgabe
14 joint targets; 100-step chunks
Veröffentlichung
8. November 2026

Einsatzbereich und Training

joint_position · continuous

14 DoF · horizon 100 · mean_std

Basismodell

Hugging Face LeRobot ACT

Verkörperungsumfang

Einzelne Verkörperung

Funktionsbereich

Action control

Datensatztyp

Offizielle Gewichte

Ausgabe

Aktionen

Pipeline-Aufgabe

robotics

Vorhersageausgabe

chunk_size=100; n_action_steps=100

Roboterform

Zweiarmplattform / Roboterarm

Aufgabenumfang

Einzelne Aufgabe

Umfang der Trainingsdaten

48,069 training frames

Quellen der Trainingsdaten

Seven simulated datasets; 119 cup episodes

Training

Imitationslernen

Abstammung

Abgeleitet von: ACT · LeRobot ACT ↗

  • Basismodell · Verkörperungsumfang · Funktionsbereich · Datensatztyp · Pipeline-Aufgabe · Roboterform · Aufgabenumfang · Umfang der Trainingsdaten · Quellen der Trainingsdaten · Training · joint_position · continuous · Abstammung · LeRobot ACT: Kite ML, Inc.
  • Ausgabe · Vorhersageausgabe · joint_position · continuous: Kite ML, Inc.

Verwandte Roboter

Alle Parameter und Prüfhinweise (Modellkarte)

Kite ML, Inc. publishes kiteml/trossen-bin-picking-sim-act-cup-all: the all-seven, 30,000-step task-trained derivative of Hugging Face LeRobot ACT. It is not a new ACT algorithm or a new weight release on 2026-11-18. No initialization from a specific pretrained ACT checkpoint is established; the exact upstream training revision is undisclosed. Trossen Robotics supplies the original recordings/robot platform, not authorship of this derivative. Other baseline, 90k and cup-real artifacts remain separate.

Training: 17 cup episodes in each of seven simulated datasets (baseline, dim, very-dim, warm, fluorescent, window, extra-objects), 119 episodes / 48,069 frames. Batch 8, one A100, 30,000 steps, about 78 minutes. Training hardware is not a minimum inference requirement. Four RGB inputs (low, high, left and right wrists), each 3×480×640, plus 14 joint positions; outputs 14 joint targets in 100-step chunks, n_obs_steps=1, n_action_steps=100. No language input. Bimanual WidowX AI context, LeRobot >=0.6.

Saved configuration: ResNet18 / ImageNet1K V1; Transformer width 512, 8 heads, feedforward 3200, ReLU, encoder 4 / decoder 1; VAE enabled, latent 32, VAE encoder 4; dropout 0.1, KL 10; MEAN_STD for images/state/action; CUDA device; AMP and PEFT disabled; learning rate 0.00001, weight decay 0.0001, backbone learning rate 0.00001. These settings are unchanged in this event.

Evaluation protocol: author-run closed-loop MuJoCo simulated station, protocol identified by report commit f61bcdc1211db23873c28aa785cb4ada8ccccaad (2026-11-17 23:39:14 UTC; Taipei 2026-11-18 07:39:14). The model-card update is 2a30bac7845f0f30b32fd87a73044aced5e3ab57, 2026-11-17 23:39:05 UTC. Sources checked 2026-11-18. Held-out source episode 4 is left-arm pickup, episode 14 right-arm pickup; each condition has 20 trials (10 per layout), cup starts within 1 cm. Success requires lifting at least 3 cm and ending inside the tray. Maximum 19 seconds, executing each 100-step chunk in full.

Current author-reported placements: room light 2/20; low light 0/20; warm light 3/20; extra objects 1/20; total 6/80 for these four conditions only. All six placements are on the left-arm layout; right-arm layout 0/40. Low light is 30–60%, warm light 2700–3500 K. Wrist-camera cables now follow the datasets' shape and arm joints, replacing fixed hooks. Test images are simulator-rendered; the low camera uses an empty real-station photograph and real-cup cutouts. This differs from the training render domain and is not evidence of generalization.

Historical protocol only: old card e2ccca8fc27cad57d2ac308970a1d023b470c847 reported room 2/20, low 2/20, warm 1/20, extra objects 0/20 with the older fixed-hook simulation. Preserve separately; never pool with current trials or interpret the change as a weight/capability gain or loss. No real-robot test or independent reproduction. These are publisher claims checked against its report and run JSON, not independently validated outcomes.

Complete weights originally uploaded 2026-11-08 23:04:42 UTC (Taipei 2026-11-09 07:04:42), commit 991c82d5d7e8eff8fd0851c1b98446b3211a4545. Only README.md changed between old and new trees. model.safetensors remains 206,765,304 bytes, LFS SHA256 51ee9bb19ccea385c3a66e83cd2c63210c38cf410081bedfe387392be73d1833. The current pinned tree is a metadata revision of that existing checkpoint. Public weights and datasets; model-card metadata declares Apache-2.0, config.license=null, no standalone LICENSE in the tree. This is the publisher's declaration, not a file-by-file legal audit. Commercial permission, full-model parameter count, action frame/units and minimum inference hardware remain unverified.

https://huggingface.co/kiteml/trossen-bin-picking-sim-act-cup-all/blob/2a30bac7845f0f30b32fd87a73044aced5e3ab57/README.md

https://huggingface.co/kiteml/trossen-bin-picking-sim-act-cup-all/blob/e2ccca8fc27cad57d2ac308970a1d023b470c847/README.md

https://huggingface.co/kiteml/trossen-bin-picking-sim-act-cup-all/blob/2a30bac7845f0f30b32fd87a73044aced5e3ab57/config.json

https://github.com/kite-ml/trossen-bin-picking-sim-datasets/blob/f61bcdc1211db23873c28aa785cb4ada8ccccaad/README.md

https://raw.githubusercontent.com/kite-ml/trossen-bin-picking-sim-datasets/f61bcdc1211db23873c28aa785cb4ada8ccccaad/policy/results/summary.json

https://huggingface.co/kiteml/trossen-bin-picking-sim-act-cup-all/commit/991c82d5d7e8eff8fd0851c1b98446b3211a4545

https://kiteml.com/

Quelle oder Korrektur einreichen