Spaces:
Running on Zero
Running on Zero
File size: 1,673 Bytes
6a5ae10 ebab5bb 6a5ae10 ebab5bb 6a5ae10 ebab5bb bbc7e3f ebab5bb | 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 | ---
title: Clef Decision Workflow
emoji: 🎼
colorFrom: pink
colorTo: yellow
sdk: gradio
sdk_version: 6.29.0
app_file: app.py
short_description: Node-based workflow for typed decisions with Cloudflare Clef
python_version: "3.12"
startup_duration_timeout: 1h
license: apache-2.0
hf_oauth: true
hf_oauth_scopes:
- inference-api
- write-repos
models:
- Cloudflare/clef
---
Visual [gr.Workflow](https://www.gradio.app/docs/gradio/workflow) demo of
[Cloudflare/clef](https://huggingface.co/Cloudflare/clef), a 27B multimodal decision model.
The canvas wires three reference nodes — **State** (text or JSON), **Questions** (a JSON schema of
typed questions: `noul` = true/false, `choice` = named options, `score` = ordered levels) and an
optional **Image** — into a bound Python function that runs Clef in a single forward pass
(no text is generated), producing a probability for every allowed option. The results land in two
output tiles: an HTML **Decision Report** and the **SystemOne Response** JSON with the answer,
confidence and probabilities for every question.
The canvas comes pre-loaded with an image example: a food photo plus a moderation schema
(what does the photo show, how appetizing is it, is it safe to publish). Clear the image node
to run text-only decisions.
Hit **Run** to execute the pipeline. Sign in with Hugging Face to run nodes on your own
inference quota. The Space owner can edit and save the workflow graph; everyone else gets a
read-only canvas. The pipeline is also exposed through the Gradio API as `/decision_report`
(see the **View API** panel or `client.view_api()`).
[Blog](https://blog.cloudflare.com/clef-decision-models)
|