hwihwalab commited on
Commit
7668811
·
verified ·
1 Parent(s): e49a9dd

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. README.md +10 -10
  2. README_KR.md +6 -6
  3. neuro_robo_bundle.zip +2 -2
README.md CHANGED
@@ -34,7 +34,7 @@ co2_eq_emissions:
34
  training_type: "biological-connectome-extraction"
35
  geographical_location: "South Korea"
36
  hardware_used: "ESP32-S3 / Arduino Uno"
37
- pretty_name: 2026 MaleCNS Fruit Fly Connectome Physical AI Foundation Model
38
  size_categories:
39
  - 10M<n<100M
40
  model-index:
@@ -44,8 +44,8 @@ model-index:
44
  type: robotics
45
  name: Physical AI Multi-Body Locomotion & Chemotaxis
46
  dataset:
47
- name: MaleCNS 2026 166.7k Connectome Graph
48
- type: connectome/malecns-2026
49
  metrics:
50
  - type: success_rate
51
  value: 100.0
@@ -67,7 +67,7 @@ model-index:
67
  name: Total Synaptic Connections
68
  ---
69
 
70
- # 🧠 MaleCNS 2026 Connectome // Physical AI Multi-Body Robotics & Embedded Control
71
 
72
  [![Language: English](https://img.shields.io/badge/Language-English-blue?style=for-the-badge)](README.md)
73
  [![Language: 한국어](https://img.shields.io/badge/Language-한국어-green?style=for-the-badge)](README_KR.md)
@@ -76,13 +76,13 @@ model-index:
76
  [![License: MIT](https://img.shields.io/badge/License-MIT-yellow?style=for-the-badge)](https://opensource.org/licenses/MIT)
77
 
78
  [![Physical AI: LeRobot Compatible](https://img.shields.io/badge/Physical%20AI-LeRobot%20Compatible-00cc66?style=for-the-badge)](https://huggingface.co/lerobot)
79
- [![Connectome: MaleCNS 2026](https://img.shields.io/badge/Connectome-MaleCNS%20166.7k%20(2026)-10b981?style=for-the-badge)](https://janelia.org)
80
  [![Latency: 5.67ms](https://img.shields.io/badge/Latency-5.67ms%20End--to--End-e11d48?style=for-the-badge)](#-5-empirical-benchmark--key-experimental-results)
81
  [![Power: 0.05W Ultra-Low](https://img.shields.io/badge/Power-0.05W%20Ultra--Low%20(14%2C000x%20Savings)-22c55e?style=for-the-badge)](#-6-computational-efficiency--green-ai-metrics)
82
  [![Multi-Body: 8 Robots](https://img.shields.io/badge/Multi--Body-8%20Robots%20·%20100%25%20Reach-8b5cf6?style=for-the-badge)](#-4-hardware--robot-platform-compatibility-matrix)
83
  [![Sim-to-Real: ESP32/C++](https://img.shields.io/badge/Sim--to--Real-ESP32%20%2F%20Arduino%20C%2B%2B-6366f1?style=for-the-badge)](./arduino_esp32_firmware.cpp)
84
 
85
- > **166,745-Neuron Whole-Brain Fruit Fly Connectome Foundation Model for Biomorphic Multi-Body Robotics & Sim-to-Real Embedded Control.**
86
  > *[ 🌐 English Documentation ](README.md) | [ 🇰🇷 한국어 매뉴얼 ](README_KR.md) | [ 🎮 Live Interactive 3D Demo ](https://huggingface.co/spaces/hwihwalab/neuro-robo-studio)*
87
 
88
  > [!TIP]
@@ -95,9 +95,9 @@ model-index:
95
 
96
  | Parameter / Metric | Specification & Empirical Result | Architecture & Domain |
97
  | :--- | :--- | :--- |
98
- | **Model Name** | `neuro-robo-connectome` | Physical AI, LeRobot, Connectome Robotics, MaleCNS |
99
- | **Biological Substrate** | Adult Male *Drosophila melanogaster* Central Nervous System | Nature 2026 MaleCNS Connectome, FlyWire, Janelia |
100
- | **Neural Scale** | **166,745 Neurons · 2,753,975 Synapses · 815 Motor Neurons** | Spiking Neural Network (SNN), LIF Neuron Dynamics |
101
  | **End-to-End Latency** | **5.67 ms** (ORN ➔ ALPN ➔ DN ➔ MN closed-loop) | Ultra-low latency, Real-time 100Hz control |
102
  | **Power Consumption** | **0.05 W** (MCU execution) vs **700 W** (Cloud GPU VLA) | Green AI, 14,000x energy efficiency, Edge AI |
103
  | **Supported Robot Bodies** | **8 Biomorphic Bodies** (CyberFly, Go1, G1, T1, MicroDuck, BH, Drone, AGV) | Hexapod, Quadruped, Bipedal Humanoid, Ornithopter, AMR |
@@ -109,7 +109,7 @@ model-index:
109
 
110
  ## 📌 1. Model Description (Overview)
111
 
112
- **MaleCNS 2026 Connectome Physical AI Foundation Model** is a whole-brain neuromorphic foundation model based on the world's largest adult male Drosophila melanogaster central nervous system (**MaleCNS 2026: 166,745 neurons, 2,753,975 synapses, 815 leg motor neurons**).
113
 
114
  Operating at **under 10ms end-to-end latency (5.67ms empirical)** and requiring **less than 0.05W of power (14,000x lower energy than cloud LLM/VLA models)**, this model drives 8 biomorphic robot bodies across diverse physical terrains for closed-loop chemotaxis pursuit and obstacle avoidance without requiring morphology-specific retraining.
115
 
 
34
  training_type: "biological-connectome-extraction"
35
  geographical_location: "South Korea"
36
  hardware_used: "ESP32-S3 / Arduino Uno"
37
+ pretty_name: "Neuro-Robo Connectome: Whole-Brain Physical AI Foundation Model (MaleCNS & FlyWire)"
38
  size_categories:
39
  - 10M<n<100M
40
  model-index:
 
44
  type: robotics
45
  name: Physical AI Multi-Body Locomotion & Chemotaxis
46
  dataset:
47
+ name: MaleCNS & FlyWire Connectome Graph
48
+ type: connectome/drosophila-dual
49
  metrics:
50
  - type: success_rate
51
  value: 100.0
 
67
  name: Total Synaptic Connections
68
  ---
69
 
70
+ # 🧠 Neuro-Robo Connectome | Whole-Brain Physical AI Foundation Model for Biomorphic Robotics
71
 
72
  [![Language: English](https://img.shields.io/badge/Language-English-blue?style=for-the-badge)](README.md)
73
  [![Language: 한국어](https://img.shields.io/badge/Language-한국어-green?style=for-the-badge)](README_KR.md)
 
76
  [![License: MIT](https://img.shields.io/badge/License-MIT-yellow?style=for-the-badge)](https://opensource.org/licenses/MIT)
77
 
78
  [![Physical AI: LeRobot Compatible](https://img.shields.io/badge/Physical%20AI-LeRobot%20Compatible-00cc66?style=for-the-badge)](https://huggingface.co/lerobot)
79
+ [![Connectome: MaleCNS & FlyWire](https://img.shields.io/badge/Connectome-MaleCNS%20(166.7k)%20%26%20FlyWire%20(139.2k)-10b981?style=for-the-badge)](https://janelia.org)
80
  [![Latency: 5.67ms](https://img.shields.io/badge/Latency-5.67ms%20End--to--End-e11d48?style=for-the-badge)](#-5-empirical-benchmark--key-experimental-results)
81
  [![Power: 0.05W Ultra-Low](https://img.shields.io/badge/Power-0.05W%20Ultra--Low%20(14%2C000x%20Savings)-22c55e?style=for-the-badge)](#-6-computational-efficiency--green-ai-metrics)
82
  [![Multi-Body: 8 Robots](https://img.shields.io/badge/Multi--Body-8%20Robots%20·%20100%25%20Reach-8b5cf6?style=for-the-badge)](#-4-hardware--robot-platform-compatibility-matrix)
83
  [![Sim-to-Real: ESP32/C++](https://img.shields.io/badge/Sim--to--Real-ESP32%20%2F%20Arduino%20C%2B%2B-6366f1?style=for-the-badge)](./arduino_esp32_firmware.cpp)
84
 
85
+ > **166,745-Neuron Whole-Brain Fruit Fly Connectome Foundation Model (MaleCNS & FlyWire FAFB) for Biomorphic Multi-Body Robotics & Sim-to-Real Embedded Control.**
86
  > *[ 🌐 English Documentation ](README.md) | [ 🇰🇷 한국어 매뉴얼 ](README_KR.md) | [ 🎮 Live Interactive 3D Demo ](https://huggingface.co/spaces/hwihwalab/neuro-robo-studio)*
87
 
88
  > [!TIP]
 
95
 
96
  | Parameter / Metric | Specification & Empirical Result | Architecture & Domain |
97
  | :--- | :--- | :--- |
98
+ | **Model Name** | `neuro-robo-connectome` | Physical AI, LeRobot, Connectome Robotics, MaleCNS & FlyWire |
99
+ | **Biological Substrate** | Adult Male (MaleCNS) & Female (FlyWire FAFB) *Drosophila* Whole CNS | Nature 2026 MaleCNS Connectome & Princeton FlyWire |
100
+ | **Neural Scale** | **166,745 (Male) / 139,255 (Female) Neurons · 2.75M~3.28M Synapses · 815 Motor Neurons** | Spiking Neural Network (SNN), LIF Neuron Dynamics |
101
  | **End-to-End Latency** | **5.67 ms** (ORN ➔ ALPN ➔ DN ➔ MN closed-loop) | Ultra-low latency, Real-time 100Hz control |
102
  | **Power Consumption** | **0.05 W** (MCU execution) vs **700 W** (Cloud GPU VLA) | Green AI, 14,000x energy efficiency, Edge AI |
103
  | **Supported Robot Bodies** | **8 Biomorphic Bodies** (CyberFly, Go1, G1, T1, MicroDuck, BH, Drone, AGV) | Hexapod, Quadruped, Bipedal Humanoid, Ornithopter, AMR |
 
109
 
110
  ## 📌 1. Model Description (Overview)
111
 
112
+ **Neuro-Robo Connectome Physical AI Foundation Model** is a whole-brain neuromorphic foundation model integrating the world's most advanced adult *Drosophila melanogaster* central nervous systems (**MaleCNS 2026: 166,745 neurons & FlyWire FAFB: 139,255 neurons, 2,753,975~3,280,000 synapses, 815 leg motor neurons**).
113
 
114
  Operating at **under 10ms end-to-end latency (5.67ms empirical)** and requiring **less than 0.05W of power (14,000x lower energy than cloud LLM/VLA models)**, this model drives 8 biomorphic robot bodies across diverse physical terrains for closed-loop chemotaxis pursuit and obstacle avoidance without requiring morphology-specific retraining.
115
 
README_KR.md CHANGED
@@ -1,4 +1,4 @@
1
- # 🧠 MaleCNS 2026 커넥톰 // 피지컬 AI 다중 생체모방 로보틱스 & 임베디드 제어 파운데이션 모델
2
 
3
  [![Language: English](https://img.shields.io/badge/Language-English-blue?style=for-the-badge)](README.md)
4
  [![Language: 한국어](https://img.shields.io/badge/Language-한국어-green?style=for-the-badge)](README_KR.md)
@@ -7,13 +7,13 @@
7
  [![License: MIT](https://img.shields.io/badge/License-MIT-yellow?style=for-the-badge)](https://opensource.org/licenses/MIT)
8
 
9
  [![Physical AI: LeRobot Compatible](https://img.shields.io/badge/Physical%20AI-LeRobot%20호환-00cc66?style=for-the-badge)](https://huggingface.co/lerobot)
10
- [![Connectome: MaleCNS 2026](https://img.shields.io/badge/Connectome-MaleCNS%20166.7k%20(2026)-10b981?style=for-the-badge)](https://janelia.org)
11
  [![Latency: 5.67ms](https://img.shields.io/badge/초저지연-5.67ms%20실시간%20루프-e11d48?style=for-the-badge)](#-5-4대-핵심-실험-실측-데이터-empirical-benchmark-results)
12
  [![Power: 0.05W Ultra-Low](https://img.shields.io/badge/초저전력-0.05W%20Ultra--Low%20(14%2C000배%20절감)-22c55e?style=for-the-badge)](#-6-초저전력-친환경-ai-연산-효율-지표-green-ai-metrics)
13
  [![Multi-Body: 8 Robots](https://img.shields.io/badge/다중바디-8종%20로봇%20·%20100%25%20도달-8b5cf6?style=for-the-badge)](#-4-하드웨어-및-로봇-플랫폼-호환성-매트릭스)
14
  [![Sim-to-Real: ESP32/C++](https://img.shields.io/badge/Sim--to--Real-ESP32%20%2F%20Arduino%20C%2B%2B-6366f1?style=for-the-badge)](./arduino_esp32_firmware.cpp)
15
 
16
- > **166,745개 전뇌 초파리 연결망 기반 다중 생체모방 로보틱스 및 Sim-to-Real 초저전력 임베디드 제어 피지컬 AI 파운데이션 모델.**
17
  > *[ 🌐 English Documentation ](README.md) | [ 🇰🇷 한국어 매뉴얼 ](README_KR.md) | [ 🎮 실시간 3D 웹 라이브 데모 ](https://huggingface.co/spaces/hwihwalab/neuro-robo-studio)*
18
 
19
  > [!TIP]
@@ -27,8 +27,8 @@
27
  | 파라미터 / 평가 항목 | 실측 사양 및 수치 (Empirical Result) | 아키텍처 및 연구 도메인 |
28
  | :--- | :--- | :--- |
29
  | **모델 공식 명칭** | `neuro-robo-connectome` | Physical AI, LeRobot, Connectome Robotics, 피지컬 AI, 뉴로모픽 |
30
- | **생체 모체 기원** | 2026 수컷 성체 초파리 중추신경계 전뇌 커넥톰 | Nature 2026 MaleCNS, FlyWire, 구글 리서치, 자넬리아 |
31
- | **신경망 규모** | **166,745개 뉴런 · 2,753,975개 시냅스 · 815개 운동 뉴런** | Spiking Neural Network (SNN), LIF 미분방정식 |
32
  | **신경 루프 지연시간** | **5.67 ms** (ORN ➔ ALPN ➔ DN ➔ MN 폐루프) | 초저지연 제어, 실시간 100Hz 모터 토크 생성 |
33
  | **전력 소모량** | **0.05 W** (MCU) vs **700 W** (클라우드 H100 GPU) | Green AI, 친환경 AI, 14,000배 전력 절감, 엣지 AI |
34
  | **제어 로봇 바디 (8종)**| **생체모방 로봇 8종** (CyberFly, Go1, G1, T1, MicroDuck, BH, Drone, AGV) | 6족 곤충, 4족 로봇개, 2족 휴머노이드, 마이크로 드론, AMR |
@@ -40,7 +40,7 @@
40
 
41
  ## 📌 1. 모델 상세 개요 (Model Description)
42
 
43
- **MaleCNS 2026 Connectome Physical AI Foundation Model**은 2026년 공개된 세계 최대 규모의 수컷 성체 초파리 중추신경계(**MaleCNS 2026: 166,745개 뉴런, 2,753,975개 시냅스, 815개 척수 다리 운동 뉴런**)를 기반으로 설계된 **전뇌 뉴로모픽 피지컬 AI 파운데이션 모델**입니다.
44
 
45
  본 모델은 클라우드 거대 언어/행동 모델(VLA/LLM) 대비 **1/14,000 이하의 초저전력(0.05W) 및 10ms 미만(실측 5.67ms)의 초저지연 폐루프 제어**를 달성하여, 휴머노이드, 4족보행 로봇개, 마이크로 드론 등 8종의 상이한 생체모방 로봇을 별도의 전이 재학습 없이도 실시간으로 자율 주행 및 장애물 회피를 수행하도록 제어합니��.
46
 
 
1
+ # 🧠 뉴로-로보 커넥톰 | 생체모방 로보틱스를 위한 전뇌 피지컬 AI 파운데이션 모델
2
 
3
  [![Language: English](https://img.shields.io/badge/Language-English-blue?style=for-the-badge)](README.md)
4
  [![Language: 한국어](https://img.shields.io/badge/Language-한국어-green?style=for-the-badge)](README_KR.md)
 
7
  [![License: MIT](https://img.shields.io/badge/License-MIT-yellow?style=for-the-badge)](https://opensource.org/licenses/MIT)
8
 
9
  [![Physical AI: LeRobot Compatible](https://img.shields.io/badge/Physical%20AI-LeRobot%20호환-00cc66?style=for-the-badge)](https://huggingface.co/lerobot)
10
+ [![Connectome: MaleCNS & FlyWire](https://img.shields.io/badge/Connectome-MaleCNS%20(166.7k)%20%26%20FlyWire%20(139.2k)-10b981?style=for-the-badge)](https://janelia.org)
11
  [![Latency: 5.67ms](https://img.shields.io/badge/초저지연-5.67ms%20실시간%20루프-e11d48?style=for-the-badge)](#-5-4대-핵심-실험-실측-데이터-empirical-benchmark-results)
12
  [![Power: 0.05W Ultra-Low](https://img.shields.io/badge/초저전력-0.05W%20Ultra--Low%20(14%2C000배%20절감)-22c55e?style=for-the-badge)](#-6-초저전력-친환경-ai-연산-효율-지표-green-ai-metrics)
13
  [![Multi-Body: 8 Robots](https://img.shields.io/badge/다중바디-8종%20로봇%20·%20100%25%20도달-8b5cf6?style=for-the-badge)](#-4-하드웨어-및-로봇-플랫폼-호환성-매트릭스)
14
  [![Sim-to-Real: ESP32/C++](https://img.shields.io/badge/Sim--to--Real-ESP32%20%2F%20Arduino%20C%2B%2B-6366f1?style=for-the-badge)](./arduino_esp32_firmware.cpp)
15
 
16
+ > **166,745개 전뇌 초파리 연결망(수컷 MaleCNS 및 암컷 FlyWire FAFB) 기반 다중 생체모방 로보틱스 및 Sim-to-Real 초저전력 임베디드 제어 피지컬 AI 파운데이션 모델.**
17
  > *[ 🌐 English Documentation ](README.md) | [ 🇰🇷 한국어 매뉴얼 ](README_KR.md) | [ 🎮 실시간 3D 웹 라이브 데모 ](https://huggingface.co/spaces/hwihwalab/neuro-robo-studio)*
18
 
19
  > [!TIP]
 
27
  | 파라미터 / 평가 항목 | 실측 사양 및 수치 (Empirical Result) | 아키텍처 및 연구 도메인 |
28
  | :--- | :--- | :--- |
29
  | **모델 공식 명칭** | `neuro-robo-connectome` | Physical AI, LeRobot, Connectome Robotics, 피지컬 AI, 뉴로모픽 |
30
+ | **생체 모체 기원** | 성체 초파리 중추신경계 전뇌 커넥톰 (수컷 MaleCNS & 암컷 FlyWire FAFB) | Nature 2026 MaleCNS, FlyWire, 구글 리서치, 자넬리아 |
31
+ | **신경망 규모** | **166,745개 (수컷) / 139,255개 (암컷) 뉴런 · 275만~328만 개 시냅스 · 815개 운동 뉴런** | Spiking Neural Network (SNN), LIF 미분방정식 |
32
  | **신경 루프 지연시간** | **5.67 ms** (ORN ➔ ALPN ➔ DN ➔ MN 폐루프) | 초저지연 제어, 실시간 100Hz 모터 토크 생성 |
33
  | **전력 소모량** | **0.05 W** (MCU) vs **700 W** (클라우드 H100 GPU) | Green AI, 친환경 AI, 14,000배 전력 절감, 엣지 AI |
34
  | **제어 로봇 바디 (8종)**| **생체모방 로봇 8종** (CyberFly, Go1, G1, T1, MicroDuck, BH, Drone, AGV) | 6족 곤충, 4족 로봇개, 2족 휴머노이드, 마이크로 드론, AMR |
 
40
 
41
  ## 📌 1. 모델 상세 개요 (Model Description)
42
 
43
+ **Neuro-Robo Connectome Physical AI Foundation Model**은 세계 최대 규모의 초파리 중추신경계 전뇌 연결망(**수컷 MaleCNS: 166,745개 뉴런 및 암컷 FlyWire FAFB: 139,255개 뉴런, 2,753,975~3,280,000개 시냅스, 815개 척수 다리 운동 뉴런**)을 기반으로 설계된 **전뇌 뉴로모픽 피지컬 AI 파운데이션 모델**입니다.
44
 
45
  본 모델은 클라우드 거대 언어/행동 모델(VLA/LLM) 대비 **1/14,000 이하의 초저전력(0.05W) 및 10ms 미만(실측 5.67ms)의 초저지연 폐루프 제어**를 달성하여, 휴머노이드, 4족보행 로봇개, 마이크로 드론 등 8종의 상이한 생체모방 로봇을 별도의 전이 재학습 없이도 실시간으로 자율 주행 및 장애물 회피를 수행하도록 제어합니��.
46
 
neuro_robo_bundle.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0f8525302ffb50b622444d1356d3e3ea985d1109cd7aaff920bbbe2b0594d49e
3
- size 12901796
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7c3d544ace203640e90ef50a467514af271b8c938ce9473e91dbba8fd990b3a0
3
+ size 12901957