Downloads · 30 days
5
22% of all-time downloads
asimov-v1-adaptive-engine
rijal028/asimov-v1-adaptive-engine
asimov-v1-adaptive-engine is a reinforcement learning model from rijal028. Use it for the reinforcement learning task on the model card, and read the license before you ship it in a product. The card lists the license as mit.
Model checkpoint adaptif mandiri (In-Flight Continual Self-Recovery) untuk robot humanoid bipedal Menlo Asimov v1 (25-DOF) berbasis Active Inference / Predictive Coding.
All-time downloads
23
Public
Repo size
2 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.pt2 MB · 100%
At a glance
- Task
- Reinforcement Learning
- License
- mit
- Access
- Public
- Created
- Aug 23, 2026
- Updated
- Aug 23, 2026
- SHA
- dccbebf0
Model card
From the Hugging Face model README
Menlo Asimov v1: On-Device Predictive Self-Recovery Engine
Model checkpoint adaptif mandiri (In-Flight Continual Self-Recovery) untuk robot humanoid bipedal Menlo Asimov v1 (25-DOF) berbasis Active Inference / Predictive Coding.
Ringkasan Performa Benchmark (20 Stage Malfungsi)
- Baseline (Static Policy): 5,818 steps
- Metode Kita (Tuned LR 1e-06): 6,267 steps (+449 steps surplus)
- Win Rate: 18 / 20 Stage (90% Win Rate)
- Parameter Drift: 0.977% (Preservasi gait stabil tanpa catastrophic forgetting)
Berkas Checkpoint:
asimov_adapted_policy.pt: Policy network yang sudah terkonsolidasi dengan pola kompensasi anomali.asimov_base_policy.pt: Nominal policy dasar.asimov_world_model.pt: Dynamics predictor untuk kalkulasi surprise & mental imagination search.config.json: Metadata dimensi observasi, aksi, dan konfigurasi pelatihan.
- Task
- Reinforcement Learning
- License
- mit
- Created
- Aug 23, 2026
- Updated
- Aug 23, 2026