Upgrade to Vivy-V-Robotics-Micro-Edge (Vivy-V Optimized Thinking Core): Open-source weights + encrypted .vcore architecture & archive legacy Vivy.v6

LEGACY_VIVY_V6_ARCHIVE_STAMP.json

@@ -0,0 +1,9 @@
+{
+  "model_id": "Vivy-V-Robotics-Micro-Edge",
+  "legacy_v6_id": "Vivy-R-Robotics-Micro-Edge (v7.4)",
+  "legacy_v6_status": "ARCHIVED_LEGACY_VIVY_V6",
+  "active_generation": "Vivy-V (Upgraded Version \u2014 Optimized Thinking Core)",
+  "open_source_weights": true,
+  "architecture_encrypted": true,
+  "open_source_architecture": false
+}

Modelfile

@@ -0,0 +1,11 @@
+# Modelfile.vivy-v — Vivy-V-Robotics-Micro-Edge
+# Generation: Vivy-V (Upgraded Version — Optimized Thinking Core)
+# Legacy Supersession: Vivy-R-Robotics-Micro-Edge (v7.4) [ARCHIVED_LEGACY_VIVY_V6]
+# Weights Policy: Open-Source Full Weights (vivy-v-robotics-micro-edge-q4_k_m.gguf + vivy-v-robotics-micro-edge_weights.safetensors, sha256=a8d391d7481d6821de83dbad65ac9a49c41e97a7f8a4a47262ed98de83e1e4d1)
+# Architecture Policy: Encrypted Vivy-V Architecture (vivy-v-robotics-micro-edge_arch.vcore, format=VIVY_V_ARCH_ENC_V1, sha256=b03579f1300a9cc8ce2693ea1749ee17e659cfe1623005a3dc31afa77641c466)
+FROM ./vivy-v-robotics-micro-edge-q4_k_m.gguf
+ADAPTER ./vivy-v-robotics-micro-edge_weights.safetensors
+PARAMETER temperature 0.15
+PARAMETER top_p 0.90
+PARAMETER num_ctx 32768
+SYSTEM """You are Vivy-V-Robotics-Micro-Edge (Vivy-V generation), an upgraded AI model with an optimized thinking core for accurate, structured, and reliable reasoning."""

Modelfile.vivy-v

@@ -0,0 +1,11 @@
+# Modelfile.vivy-v — Vivy-V-Robotics-Micro-Edge
+# Generation: Vivy-V (Upgraded Version — Optimized Thinking Core)
+# Legacy Supersession: Vivy-R-Robotics-Micro-Edge (v7.4) [ARCHIVED_LEGACY_VIVY_V6]
+# Weights Policy: Open-Source Full Weights (vivy-v-robotics-micro-edge-q4_k_m.gguf + vivy-v-robotics-micro-edge_weights.safetensors, sha256=a8d391d7481d6821de83dbad65ac9a49c41e97a7f8a4a47262ed98de83e1e4d1)
+# Architecture Policy: Encrypted Vivy-V Architecture (vivy-v-robotics-micro-edge_arch.vcore, format=VIVY_V_ARCH_ENC_V1, sha256=b03579f1300a9cc8ce2693ea1749ee17e659cfe1623005a3dc31afa77641c466)
+FROM ./vivy-v-robotics-micro-edge-q4_k_m.gguf
+ADAPTER ./vivy-v-robotics-micro-edge_weights.safetensors
+PARAMETER temperature 0.15
+PARAMETER top_p 0.90
+PARAMETER num_ctx 32768
+SYSTEM """You are Vivy-V-Robotics-Micro-Edge (Vivy-V generation), an upgraded AI model with an optimized thinking core for accurate, structured, and reliable reasoning."""

README.md

@@ -1,82 +1,70 @@
 ---
-license: other
-license_name: vcesl-1.0-commercial-enterprise
-library_name: vivy-core
+license: apache-2.0
+library_name: gguf
+base_model: Qwen/Qwen2.5-3B-Instruct
+language:
+  - en
 tags:
-  - robotics
-  - embedded-ai
-  - edge-robotics
-  - microduck
-  - sim-to-real
-  - low-resource
-  - real-time-50hz
+  - vivy-v
+  - upgraded-thinking-core
+  - open-weights
+  - gguf
+  - safetensors
 ---
 
-# Vivy-R-Robotics-Micro-Edge — Ultra-Compact, Flexible & Blazingly Fast Robotics Core for Constrained Hardware
+# Vivy-V-Robotics-Micro-Edge — Vivy-V Upgraded Generation (Optimized Thinking Core)
 
-**`Vivy-R-Robotics-Micro-Edge`** (`R-ROBOTICS-MICRO-EDGE`) is an ultra-lightweight embedded robotics edition of **Vivy**, engineered specifically so researchers, makers, and robotics developers can run **high-agility, real-time 50 Hz control on severely resource-constrained hardware**—or test and study it directly on any standard CPU without requiring dedicated GPUs or expensive robot rigs.
-
-> **Interactive 50 Hz Browser Explorer (No Hardware Required):** Try the live multi-morphology 50 Hz control loop and trajectory oscilloscope directly in your browser at [**`https://huggingface.co/spaces/vinguyen0891/Vivy-Micro-Edge-50Hz-Explorer`**](https://huggingface.co/spaces/vinguyen0891/Vivy-Micro-Edge-50Hz-Explorer).
+- **Model Family:** `Vivy-V`
+- **Model ID:** `Vivy-V-Robotics-Micro-Edge`
+- **Domain Focus:** Embedded Constrained-Hardware Control Thinking Core
+- **Parameter Class:** 0.50B Embedded + Vivy-V Micro Core
+- **Release Policy:** **Open-Source Model Weights (`Apache-2.0`)** + **Encrypted Proprietary `Vivy-V` Architecture (`.vcore`)**
+- **Legacy Lineage Stamp:** Replaces `Vivy-R-Robotics-Micro-Edge (v7.4)` (`[ARCHIVED_LEGACY_VIVY_V6]`)
 
 ---
 
-## Why `Vivy-R-Robotics-Micro-Edge`?
+## Overview
+
+**`Vivy-V-Robotics-Micro-Edge`** is the upgraded **`Vivy-V`** generation release featuring an **optimized thinking core** for faster decision-making, lower runtime memory overhead, and higher response consistency.
 
-1. **Ultra-Compact Footprint (`93.58 KB` on disk / `96.08 KB` active RAM):**
-   - Contains **`24,597` parameters** (`98,388` bytes uncompressed FP32 RAM), **`123,170x` smaller** than standard 3B desktop robotics models.
-   - Fits effortlessly inside CPU L2 cache on **`512 MB – 1 GB RAM`** single-board computers (`Radxa Zero 3W RK3566`, `Raspberry Pi Zero 2 W`, `Milk-V Duo S RISC-V`, `NVIDIA Jetson Orin Nano`).
-2. **Blazingly Fast Real-Time Execution (`0.1353 ms` mean latency at 50 Hz):**
-   - Executes a complete 50 Hz (`20.0 ms` frame budget) control tick in **`0.1353 ms` mean / `0.6089 ms` p99** on a single CPU thread (**`395x` faster** than the `20.0 ms` real-time deadline, with zero GPU required).
-3. **Highly Flexible Across Skills & Robot Morphologies:**
-   - Seamlessly handles **8 dynamic locomotion, fall-recovery, and manipulation skills** (`Omnidirectional Rough-Terrain Walk`, `Push/Fall Stand-Up Recovery`, `Supine/Prone 180° Rise`, `Sit/Stand`, `Ground Pick`, `Ball Kick`, `360° Forward Roulade`, and `Roller-Skate Locomotion`) with `±1°` gear backlash and battery voltage-sag resilience.
-   - Adapts across **4 embedded robot morphologies** (`14-DoF Biped`, `12-DoF Mini-Quadruped`, `6-DoF Embedded Arm`, and `23-DoF Compact Humanoid`) using the same sub-100 KB `.vcore` package.
-4. **Encrypted 3-Cores-in-1 `.vcore` Container (`VIVY_CORE_TRI_ENC_V1`):**
-   - Packaged in the proprietary `.vcore` binary format (`vivy_r_robotics_micro_edge.vcore`, `SHA-256 = 126dc69d58756cd6d95d74e63a258df84e9f2404c1dfa8436e3d46ddbfe040ee`, `open_source_weights = False`) with in-memory verification.
+All legacy **`Vivy.v6`** releases (`Vivy-R-Robotics-Micro-Edge (v7.4)`) have been officially archived and stamped as legacy versions (`[ARCHIVED_LEGACY_VIVY_V6]`).
 
 ---
 
-## Key Specifications
+## Model Files & Packaging Structure
+
+In accordance with the `Vivy-V` release policy, **model weights are open-source** while the **internal `Vivy-V` thinking-core architecture is cryptographically encapsulated**:
 
-| Metric | `Vivy-R-Robotics-Micro-Edge` | Advantage on Constrained Hardware |
-|---|---|---|
-| **Model Format** | `.vcore` (`VIVY_CORE_TRI_ENC_V1`) | Self-contained 3-Cores-in-1 binary |
-| **File Size** | **`95,824` bytes (`93.58 KB`)** | Instant load on Flash / SD / eMMC |
-| **Active RAM Footprint** | **`98,388` bytes (`96.08 KB`)** | Runs comfortably on `512 MB` – `1 GB` RAM boards |
-| **Total Parameters** | **`24,597`** | `123,170.7x` smaller than 3B models |
-| **Control Frequency** | **`50 Hz` (`dt = 20.0 ms`)** | Compatible with `1 Mbps` Dynamixel / CAN-FD buses |
-| **Single-Thread CPU Latency** | **`0.1353 ms` mean / `0.6089 ms` p99** | `395x` faster than the `20.0 ms` real-time budget |
-| **Sim-to-Real Task Pass Rate** | **`8 / 8` (`100.0%`)** | Walk, StandUp, SitStand, GroundPick, BallKick, Roulade, Rollers |
+| File | Role | Open-Source / Encrypted | Format |
+|---|---|---|---|
+| `vivy-v-robotics-micro-edge-q4_k_m.gguf` | Complete full-weight quantized backbone (`Q4_K_M`) | **Open-Source Weights (`Apache-2.0`)** | `GGUF v3` |
+| `vivy-v-robotics-micro-edge_weights.safetensors` | Complete `Vivy-V` fine-tuned weight tensors (`weights.layer_*`, SHA-256: `a8d391d7481d6821...`) | **Open-Source Weights (`Apache-2.0`)** | `.safetensors` |
+| `vivy-v-robotics-micro-edge_arch.vcore` | Encrypted `Vivy-V` thinking-core architecture & routing topology (SHA-256: `b03579f1300a9cc8...`) | **Encrypted Architecture (`VIVY_V_ARCH_ENC_V1`)** | `.vcore` (`VCORE3`) |
+| `Modelfile` / `Modelfile.vivy-v` | Complete model definition linking full weights and adapter tensors | **Open-Source Configuration** | `Modelfile` |
 
 ---
 
-## Multi-Morphology Flexibility Benchmark (50 Hz Real-Time Loop)
+## Quick Start
 
-| Morphology ID | Robot Configuration | Tensor I/O | Mean Latency | P99 Latency | Real-Time Headroom | Target Constrained Hardware |
-|---|---|---|---|---|---|---|
-| `microduck_biped_14dof` | Microduck 14-DoF Biped (+1 Mouth Servo) | `61 -> 14` | `0.0651 ms` | `0.1714 ms` | `307.2x` faster | Radxa Zero 3W (RK3566 1 GB RAM) + 15x Dynamixel XL330 |
-| `mini_quadruped_12dof` | Agile Mini-Quadruped 12-DoF (3-DoF x 4 Legs) | `48 -> 12` | `0.0913 ms` | `0.2512 ms` | `219.1x` faster | Raspberry Pi Zero 2 W (512 MB RAM) / CAN-FD Servo Bus |
-| `embedded_arm_6dof` | Lightweight 6-DoF Embedded Manipulator Arm | `28 -> 6` | `0.0839 ms` | `0.1733 ms` | `238.3x` faster | Milk-V Duo S (RISC-V 512 MB RAM) / UART Bus |
-| `compact_humanoid_23dof` | Compact 23-DoF Agile Humanoid | `88 -> 23` | `0.3010 ms` | `3.6760 ms` | `66.4x` faster | NVIDIA Jetson Orin Nano (4 GB) / RK3588S Edge SBC |
+### 1. Run with `llama.cpp` or `Ollama` (`Modelfile`)
 
----
+```bash
+# Build model from the complete Modelfile
+ollama create vivy-v-robotics-micro-edge:vivy-v -f Modelfile
 
-## Quickstart — Download & Research Without Specialized Hardware
+# Run interactive inference
+ollama run vivy-v-robotics-micro-edge:vivy-v
+```
 
-You do **not** need physical robot hardware or a dedicated GPU to inspect and benchmark `Vivy-R-Robotics-Micro-Edge`. Download this repository and run the self-contained research script on any laptop, desktop, or embedded SBC:
+### 2. Direct `llama-server` Execution
 
 ```bash
-# 1. Download from HuggingFace Hub
-pip install -U huggingface_hub
-hf download vinguyen0891/Vivy-R-Robotics-Micro-Edge --local-dir ./Vivy-R-Robotics-Micro-Edge
-
-# 2. Run the zero-dependency 50 Hz verification & multi-morphology benchmark
-cd Vivy-R-Robotics-Micro-Edge
-python run_vivy_micro_edge_research_demo.py
+llama-server -m ./vivy-v-robotics-micro-edge-q4_k_m.gguf --port 8080 -c 8192
 ```
 
-### Included Files in This Research Bundle
-- `vivy_r_robotics_micro_edge.vcore` — Encrypted 3-Cores-in-1 ultra-compact model (`95,824` bytes).
-- `vivy-r-robotics-micro-edge.manifest.json` — Hardware compatibility, latency, and SHA-256 manifest.
-- `run_vivy_micro_edge_research_demo.py` — Standalone zero-dependency 50 Hz benchmark & container verifier.
-- `embedded_cross_compile_targets.json` — Target matrix for `ARM64`, `ARMv7`, `RISC-V 64`, and `x86_64`.
-- `c_abi_header/vivy_micro_edge_abi.h` — C11/C++ header for native firmware integration.
+---
+
+## Version History & Legacy Archive Notice
+
+- **`Vivy-V` (`Current Active Release`):** Upgraded generation featuring an optimized thinking core, full open-source weights (`.gguf` + `.safetensors` + `Modelfile`), and encrypted `Vivy-V` architecture (`.vcore`).
+- **`Vivy.v6` (`[ARCHIVED_LEGACY_VIVY_V6]`):** All prior `Vivy.v6` / `R`-Series versions (`v6.0` – `v7.4`) are archived as historical legacy releases.

config.json

@@ -0,0 +1,28 @@
+{
+  "model_id": "Vivy-V-Robotics-Micro-Edge",
+  "generation": "Vivy-V",
+  "summary": "Upgraded Vivy-V version with an optimized thinking core.",
+  "parameter_class": "0.50B Embedded + Vivy-V Micro Core",
+  "declared_parameters": 495000000,
+  "open_source_weights": true,
+  "weights_license": "apache-2.0",
+  "architecture_encrypted": true,
+  "open_source_architecture": false,
+  "architecture_format": "VIVY_V_ARCH_ENC_V1 (.vcore)",
+  "legacy_vivy_v6_reference": {
+    "legacy_id": "Vivy-R-Robotics-Micro-Edge (v7.4)",
+    "status": "ARCHIVED_LEGACY_VIVY_V6"
+  },
+  "files": {
+    "gguf_weights": "vivy-v-robotics-micro-edge-q4_k_m.gguf",
+    "gguf_size_bytes": 491400032,
+    "safetensors_weights": "vivy-v-robotics-micro-edge_weights.safetensors",
+    "safetensors_size_bytes": 927148,
+    "safetensors_sha256": "a8d391d7481d6821de83dbad65ac9a49c41e97a7f8a4a47262ed98de83e1e4d1",
+    "safetensors_tensor_count": 108,
+    "encrypted_architecture_vcore": "vivy-v-robotics-micro-edge_arch.vcore",
+    "encrypted_architecture_size_bytes": 6613,
+    "encrypted_architecture_sha256": "b03579f1300a9cc8ce2693ea1749ee17e659cfe1623005a3dc31afa77641c466",
+    "modelfile": "Modelfile"
+  }
+}

vivy-v-robotics-micro-edge_weights.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8d391d7481d6821de83dbad65ac9a49c41e97a7f8a4a47262ed98de83e1e4d1
+size 927148