[BOUNTY: 15,000 WATT] Raspberry Pi inference relay

[WattCoin-Org]

Description

Run a small LLM on Raspberry Pi and charge WATT per query via API.

Requirements

• Run lightweight model (TinyLlama, Phi-2, or similar)
• API endpoint that accepts prompt + returns model response
• Queue management for requests
• Usage tracking and stats
• Optimized for Pi 4/5 (4GB+ RAM)
• Configurable model path and parameters

Acceptance Criteria

• Model runs on Pi with acceptable latency (<30s for short prompts)
• API endpoint functional
• Usage stats tracked
• Install script included
• Tests included and passing

Bounty

15,000 WATT

Notes

• Payment verification integration is optional — focus on the inference relay working correctly.
• Include your personal Solana wallet in the PR body (not the project wallet). Format: **Payout Wallet**: <your_address>

[chiang881]

Implemented a minimal mergeable inference-relay scaffold in PR #222: https://github.com/WattCoin-Org/wattcoin/pull/222\n\nIncludes:\n- Raspberry Pi relay runner scaffold ()\n- Signed report submission flow ()\n- install script + quickstart docs\n- focused tests for result payload + signature stability

[chiang881]

Correction: PR #222 adds pi/inference_relay/relay.py, signed submission to /api/v1/inference/report, install script/docs, and tests.

[zhanglin2603]

Hi team, I read through #18 and I can own this.

• My read: this is primarily a UI/UX task tied to "[BOUNTY: 15,000 WATT] Raspberry Pi inference relay".
• I will focus on these issue-specific points first: bounty, watt, raspberry.
• First I will reproduce and confirm expected behavior for the UI/UX path.
• Then I will implement the fix with targeted tests and keep scope tight (good confidence).
• I will share a working PR update fast (24h for first patch, 48h for final polish) with verification notes.
  If this approach works for you, I can start now.