yolo-detection-2026-coral-tpu-win-wsl
starsforks
How to install yolo-detection-2026-coral-tpu-win-wsl?
sharpai/deepcamera/yolo-detection-2026-coral-tpu-win-wsl
$npx -y skills add SharpAI/DeepCamera --skill yolo-detection-2026-coral-tpu-win-wslGoogle Coral Edge TPU — real-time object detection natively via Windows WSL
SKILL.md
| 1 | # Coral TPU Object Detection (Windows WSL) |
| 2 | |
| 3 | Real-time object detection natively utilizing the Google Coral Edge TPU accelerator on your local hardware via Windows Subsystem for Linux (WSL). Detects 80 COCO classes (person, car, dog, cat, etc.) with ~4ms inference on 320x320 input. |
| 4 | |
| 5 | ## Requirements |
| 6 | |
| 7 | - **Google Coral USB Accelerator** (USB 3.0 port recommended) |
| 8 | - **WSL2** installed and running on Windows |
| 9 | - `usbipd-win` installed on the Windows host |
| 10 | |
| 11 | ## How It Works |
| 12 | |
| 13 | ``` |
| 14 | ┌─────────────────────────────────────────────────────┐ |
| 15 | │ Host (Aegis-AI on Windows) │ |
| 16 | │ frame.jpg → /tmp/aegis_detection/ │ |
| 17 | │ stdin ──→ ┌──────────────────────────────┐ │ |
| 18 | │ │ WSL Container / Environment │ │ |
| 19 | │ │ detect.py │ │ |
| 20 | │ │ ├─ loads _edgetpu.tflite │ │ |
| 21 | │ │ ├─ reads frame from disk │ │ |
| 22 | │ │ └─ runs inference on TPU │ │ |
| 23 | │ stdout ←── │ → JSONL detections │ │ |
| 24 | │ └──────────────────────────────┘ │ |
| 25 | │ USB ──→ usbipd-win bridge to WSL │ |
| 26 | └─────────────────────────────────────────────────────┘ |
| 27 | ``` |
| 28 | |
| 29 | 1. Aegis writes camera frame JPEG to shared `/tmp/aegis_detection/` workspace |
| 30 | 2. Sends `frame` event via stdin JSONL to the WSL Python instance |
| 31 | 3. `detect.py` invokes PyCoral and executes natively on the mapped USB Edge TPU inside Linux |
| 32 | 4. Returns `detections` event via stdout JSONL back to Windows Host |
| 33 | |
| 34 | ## Performance |
| 35 | |
| 36 | | Input Size | Inference | On-chip | Notes | |
| 37 | |-----------|-----------|---------|-------| |
| 38 | | 320x320 | ~4ms | 100% | Fully on TPU, best for real-time | |
| 39 | | 640x640 | ~20ms | Partial | Some layers on CPU (model segmented) | |
| 40 | |
| 41 | > **Cooling**: The USB Accelerator aluminum case acts as a heatsink. If too hot to touch during continuous inference, it will thermal-throttle. Consider active cooling or `clock_speed: standard`. |
| 42 | |
| 43 | ## Installation |
| 44 | |
| 45 | ### Windows (WSL) |
| 46 | Run `deploy.bat` — this will: |
| 47 | 1. Verify `usbipd` is installed and bind the `18d1:9302` and `1a6e:089a` Edge TPU hardware IDs. |
| 48 | 2. Setup a Python virtual environment exclusively within WSL. |
| 49 | 3. Install the Edge TPU libraries and dependencies within the WSL boundary. |
| 50 | 4. Auto-attach the device using `usbipd` seamlessly during invocation. |
People who liked this also viewed
6 picksType
- find-skillsHelps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skSkillsJul 2026
- grill-meA relentless interview to sharpen a plan or design.SkillsJul 2026
- vercel-react-best-practicesVercel's official collection of agent skillsSkillsJul 2026
- improve-codebase-architectureScan a codebase for deepening opportunities, present them as a visual HTML report, then grill through whichever one you pick.SkillsJul 2026
- setup-matt-pocock-skillsConfigure this repo for the engineering skills — set up its issue tracker, triage label vocabulary, and domain doc layout. Run once before first use of the other engineering skills.SkillsJul 2026
- triageMove issues and external PRs through a state machine of triage roles — categorise, verify, grill if needed, and write agent-ready briefs.SkillsJul 2026
