Spaces:
Running
Running
File size: 1,333 Bytes
2498dad be6685b 2498dad be6685b 2498dad be6685b |
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 |
---
title: CodeBook
emoji: π
colorFrom: purple
colorTo: red
sdk: gradio
sdk_version: 5.49.1
app_file: app.py
pinned: true
short_description: Advanced AI workflowing system.
---
# π€ CodeBook AI Workflow Space: Code or Book Generator with Self-Critique
This Hugging Face Space refines an **idea** into either:
- A *complete code project* (pseudocode β snippet β full code), or
- A *structured story* (Start β Development β Climax β Conclusion β End).
After creation, **two AI critics** rate and explain the output, and the system automatically refines it based on their feedback β a full *AI self-improvement loop*.
## π§© Workflow
1. Input your **idea** and choose **Code mode** or **Book mode**.
2. The AI refines your prompt.
3. It builds content step-by-step.
4. Two critics analyze and rate it.
5. The main model refines the output using that feedback.
## π οΈ Models Used
- Prompt refiner: `mistralai/Mixtral-8x7B-Instruct-v0.1`
- Code generator: `codellama/CodeLlama-7b-Instruct-hf`
- Book generator: `mistralai/Mixtral-8x7B-Instruct-v0.1`
- Critics: `google/gemma-2-9b-it`, `meta-llama/Meta-Llama-3-8B-Instruct`
You can modify these to lighter or open-source models depending on performance limits.
## π Deployment
Push this repo to Hugging Face as a Space and select βGradioβ as the SDK. |