Spaces:
Sleeping
Sleeping
Update README.md
Browse files
README.md
CHANGED
|
@@ -11,4 +11,72 @@ license: apache-2.0
|
|
| 11 |
short_description: Bulldog Children's Story Generation
|
| 12 |
---
|
| 13 |
|
| 14 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
short_description: Bulldog Children's Story Generation
|
| 12 |
---
|
| 13 |
|
| 14 |
+
# Tech Tales: AI-Powered Story Creation 📚
|
| 15 |
+
|
| 16 |
+
A creative storytelling application that combines landscape generation, story writing, and image generation to create illustrated children's stories featuring Champ, an adventurous English bulldog.
|
| 17 |
+
|
| 18 |
+
## 🌟 Features
|
| 19 |
+
|
| 20 |
+
### 1. Landscape Generation
|
| 21 |
+
- Automatically generates beautiful landscape scenes using Stable Diffusion
|
| 22 |
+
- Creates unique backgrounds for story inspiration
|
| 23 |
+
|
| 24 |
+
### 2. Scene Analysis
|
| 25 |
+
- Analyzes generated landscapes using SmolVLM
|
| 26 |
+
- Provides detailed descriptions of the scene elements
|
| 27 |
+
|
| 28 |
+
### 3. Story Generation
|
| 29 |
+
- Creates engaging children's stories using SmolLM2
|
| 30 |
+
- Features Champ, a lovable English bulldog as the main character
|
| 31 |
+
- Incorporates themes of confidence, teamwork, caring, and hope
|
| 32 |
+
- Ensures age-appropriate content with moral lessons
|
| 33 |
+
|
| 34 |
+
### 4. Scene Visualization
|
| 35 |
+
- Generates custom illustrations using a fine-tuned Stable Diffusion model with bulldog LoRA
|
| 36 |
+
- Creates watercolor-style images that match the story's narrative
|
| 37 |
+
- Combines text and images for a complete storybook experience
|
| 38 |
+
|
| 39 |
+
### 5. Audio Narration
|
| 40 |
+
- Converts the story into spoken narration
|
| 41 |
+
- Uses high-quality text-to-speech for natural-sounding storytelling
|
| 42 |
+
|
| 43 |
+
## 🎯 How to Use
|
| 44 |
+
|
| 45 |
+
1. Click "Generate Random Landscape" to create a starting scene
|
| 46 |
+
2. Get a brief description of the generated landscape
|
| 47 |
+
3. Generate a children's story based on the scene
|
| 48 |
+
4. Create scene prompts from the story
|
| 49 |
+
5. Generate illustrations for each scene
|
| 50 |
+
6. Add text overlays to create storybook pages
|
| 51 |
+
7. Listen to the story being read aloud
|
| 52 |
+
|
| 53 |
+
## 🛠️ Technical Details
|
| 54 |
+
|
| 55 |
+
This space uses several AI models:
|
| 56 |
+
- Stable Diffusion v1.5 for landscape generation
|
| 57 |
+
- Custom LoRA adapter for bulldog illustrations
|
| 58 |
+
- SmolVLM-500M for image analysis
|
| 59 |
+
- SmolLM2-1.7B for story generation
|
| 60 |
+
- Kokoro TTS for audio narration
|
| 61 |
+
|
| 62 |
+
## 🎨 Custom Theme
|
| 63 |
+
|
| 64 |
+
The interface features a custom color scheme with:
|
| 65 |
+
- Light blue primary elements
|
| 66 |
+
- Light red accent colors
|
| 67 |
+
- Clean, intuitive layout for easy navigation
|
| 68 |
+
|
| 69 |
+
## 📝 Notes
|
| 70 |
+
|
| 71 |
+
- Story generation takes around 30-60 seconds
|
| 72 |
+
- Each scene illustration takes approximately 30 seconds to generate
|
| 73 |
+
- The complete process typically takes 3-5 minutes
|
| 74 |
+
|
| 75 |
+
## 🙏 Acknowledgments
|
| 76 |
+
|
| 77 |
+
- Stable Diffusion by Runway & CompVis
|
| 78 |
+
- LoRA bulldog adaptation by Prof-Hunt
|
| 79 |
+
- Hugging Face for model hosting and Spaces platform
|
| 80 |
+
|
| 81 |
+
---
|
| 82 |
+
Created with ❤️ using Gradio and Hugging Face Spaces
|