Spaces:
Sleeping
Sleeping
File size: 1,407 Bytes
3f3c728 523550b 3f3c728 53f9c70 3f3c728 523550b 53f9c70 523550b 53f9c70 523550b 53f9c70 523550b |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 |
---
title: Hannah-Pilot-Interface
duplicate_from: null
sdk: docker
app_port: 7860
---
# Hannah Pilot Interface (Docker Space)
This HuggingFace Space serves a FastAPI backend for the Hannah web UI.
## ✅ Endpoints
- `GET /api/models` – lists available `.gguf` models in the repo
- `GET /api/status` – lightweight health/status (RAM + CPU)
- `POST /api/gen_title` – generates a short chat title
- `POST /api/chat` – **streams NDJSON** responses (`application/x-ndjson`)
## Frontend Setup
In your HTML frontend, set:
```js
const API_BASE = "https://fugthchat-hannah-pilot-interface.hf.space";
```
Your frontend should call:
- `POST ${API_BASE}/api/chat`
- `GET ${API_BASE}/api/models`
- `GET ${API_BASE}/api/status`
## Models
Upload your model files into the Space repo root:
- `qwen2.5-0.5b-instruct-q2_k.gguf` (Light)
- `qwen2.5-0.5b-instruct-q4_k_m.gguf` (Heavy)
The backend auto-detects all `*.gguf` files.
## Avoiding Build Loops
This Space uses:
- Docker listening on `0.0.0.0:7860`
- `app_port: 7860` in the README header
If your Space keeps restarting:
1. Confirm the app listens on port `7860`.
2. Confirm `CMD` runs `uvicorn app:app --host 0.0.0.0 --port 7860`.
3. Check the Space logs for dependency install failures.
## Notes
- This is CPU-friendly config (`n_threads=2`) for HuggingFace free tier.
- Response streaming is NDJSON (one JSON per line: `{ "text": "..." }`).
|