Spaces:
Running
Running
|
Download README.md from switchboardllm/switchboard: direct link, hf CLI and curl.
- Browser
- Download file 937 Bytes
-
https://huggingface.co/spaces/switchboardllm/switchboard/resolve/main/README.md
- Command line
-
hf download hf://spaces/switchboardllm/switchboard/README.md
-
curl -L -o README.md https://huggingface.co/spaces/switchboardllm/switchboard/resolve/main/README.md
937 Bytes
| title: Switchboard | |
| emoji: 🔀 | |
| colorFrom: yellow | |
| colorTo: red | |
| sdk: static | |
| app_file: index.html | |
| pinned: true | |
| license: apache-2.0 | |
| short_description: A 0.5B router that picks the cheapest model for each prompt | |
| models: | |
| - switchboardllm/switchboard-router-0.5b-ONNX | |
| - switchboardllm/switchboard-router-0.5b | |
| - switchboardllm/switchboard-router-0.5b-GGUF | |
| datasets: | |
| - routellm/gpt4_dataset | |
| tags: | |
| - router | |
| - llm-routing | |
| - transformers.js | |
| - webgpu | |
| # Switchboard demo | |
| Paste a prompt and see which model tier Switchboard picks: small, mid or frontier. The router ([switchboard-router-0.5b-ONNX](https://huggingface.co/switchboardllm/switchboard-router-0.5b-ONNX)) runs in your browser with Transformers.js, so prompts never leave your device. The first route downloads the model once (660 MB). | |
| Model card and results: [switchboardllm/switchboard-router-0.5b](https://huggingface.co/switchboardllm/switchboard-router-0.5b) | |