Spaces:
Paused
Paused
|
Download README.md from viarl199020/BITNET: direct link, hf CLI and curl.
- Browser
- Download file 768 Bytes
-
https://huggingface.co/spaces/viarl199020/BITNET/resolve/main/README.md
- Command line
-
hf download hf://spaces/viarl199020/BITNET/README.md
-
curl -L -o README.md https://huggingface.co/spaces/viarl199020/BITNET/resolve/main/README.md
768 Bytes
| title: Omni-Nexus Worker A | |
| emoji: 💬 | |
| colorFrom: indigo | |
| colorTo: purple | |
| sdk: docker | |
| app_port: 7860 | |
| pinned: false | |
| # Omni-Nexus · Worker de chat | |
| Llama 3.2 3B (Q4_K_M) servido con API compatible con OpenAI. | |
| Usa los binarios **oficiales precompilados** de llama.cpp: build de 3-5 min, | |
| sin compilar C++. | |
| ## Por qué ya no se usa BitNet | |
| bitnet.cpp no incluye servidor HTTP — su submódulo solo compila `llama-cli`. | |
| El target `llama-server` no existe. Y `setup_env.py` agota la RAM del builder | |
| de HF al convertir el modelo. No es parcheable. | |
| ## Uso | |
| `Odysseus → Settings → Add Local Models (Endpoint)`: | |
| ``` | |
| https://TU-USUARIO-ESTE-SPACE.hf.space/v1 | |
| ``` | |
| ## Variables | |
| | Variable | Valor | | |
| |---|---| | |
| | `WORKER_ID` | `A` | | |
| | `CHAT_TEMPLATE` | `llama3` | | |