Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -1,99 +1,54 @@
|
|
| 1 |
---
|
| 2 |
title: n8n Workflow Generator
|
| 3 |
emoji: π
|
| 4 |
-
colorFrom:
|
| 5 |
-
colorTo:
|
| 6 |
sdk: gradio
|
| 7 |
-
sdk_version: 4.
|
| 8 |
app_file: app.py
|
| 9 |
-
pinned:
|
| 10 |
license: apache-2.0
|
| 11 |
---
|
| 12 |
|
| 13 |
-
# n8n Workflow Generator
|
| 14 |
|
| 15 |
-
Generate n8n workflows using natural language!
|
| 16 |
|
| 17 |
-
## π―
|
| 18 |
|
| 19 |
-
- **
|
| 20 |
-
- **
|
| 21 |
-
- **
|
|
|
|
| 22 |
|
| 23 |
-
##
|
| 24 |
|
| 25 |
-
- **
|
| 26 |
-
- **
|
| 27 |
-
- **
|
| 28 |
-
- **n8n JSON Export:** Import directly into your n8n instance
|
| 29 |
-
- **Adjustable Parameters:** Control creativity and output length
|
| 30 |
|
| 31 |
-
##
|
| 32 |
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
| Error Handling | 80% |
|
| 38 |
-
| Loops | 67% |
|
| 39 |
-
| Branching | 67% |
|
| 40 |
-
| **Overall** | **92.4%** |
|
| 41 |
|
| 42 |
-
##
|
| 43 |
-
|
| 44 |
-
Try these prompts to see what the model can do:
|
| 45 |
|
| 46 |
- "Create a webhook that sends data to Slack"
|
| 47 |
-
- "
|
| 48 |
-
- "Webhook receives
|
| 49 |
-
- "
|
| 50 |
-
- "Fetch GitHub issues, if priority is high send to Slack, else email"
|
| 51 |
-
|
| 52 |
-
## π οΈ How It Works
|
| 53 |
-
|
| 54 |
-
1. **Input:** You describe your workflow in natural language
|
| 55 |
-
2. **Generation:** Fine-tuned Qwen2.5-Coder-1.5B generates TypeScript DSL
|
| 56 |
-
3. **Conversion:** Code is converted to n8n JSON format
|
| 57 |
-
4. **Visualization:** Workflow structure is displayed visually
|
| 58 |
-
5. **Export:** Copy and import into your n8n instance
|
| 59 |
-
|
| 60 |
-
## π Model Details
|
| 61 |
-
|
| 62 |
-
- **Base Model:** Qwen/Qwen2.5-Coder-1.5B-Instruct
|
| 63 |
-
- **Fine-tuning Method:** LoRA (Low-Rank Adaptation)
|
| 64 |
-
- **Dataset:** Curated n8n workflows from GitHub
|
| 65 |
-
- **Training Framework:** Transformers + PEFT
|
| 66 |
-
|
| 67 |
-
## π Training Details
|
| 68 |
-
|
| 69 |
-
- **LoRA Rank:** 16
|
| 70 |
-
- **LoRA Alpha:** 32
|
| 71 |
-
- **Learning Rate:** 2e-4
|
| 72 |
-
- **Training Strategy:** Early stopping with validation monitoring
|
| 73 |
-
- **Hardware:** NVIDIA Tesla T4 GPU
|
| 74 |
-
|
| 75 |
-
## π Usage Tips
|
| 76 |
-
|
| 77 |
-
- **Be specific:** More details = better results
|
| 78 |
-
- **Use n8n terminology:** Mention specific nodes like "webhook", "Slack", "HTTP Request"
|
| 79 |
-
- **Describe the flow:** "When X happens, do Y, then Z"
|
| 80 |
-
- **Adjust temperature:** Lower (0.1-0.3) for consistency, higher (0.5-0.8) for creativity
|
| 81 |
-
|
| 82 |
-
## π§ Limitations
|
| 83 |
-
|
| 84 |
-
- Works best with common n8n patterns
|
| 85 |
-
- May struggle with very complex branching (>5 conditions)
|
| 86 |
-
- Advanced error handling might need manual refinement
|
| 87 |
-
- Custom node configurations may need adjustment
|
| 88 |
|
| 89 |
-
##
|
| 90 |
|
| 91 |
-
|
|
|
|
| 92 |
|
| 93 |
## π Acknowledgments
|
| 94 |
|
| 95 |
Built with:
|
| 96 |
- [Qwen2.5-Coder](https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B-Instruct)
|
|
|
|
| 97 |
- [Hugging Face Transformers](https://github.com/huggingface/transformers)
|
| 98 |
-
- [PEFT](https://github.com/huggingface/peft)
|
| 99 |
-
- [Gradio](https://gradio.app)
|
|
|
|
| 1 |
---
|
| 2 |
title: n8n Workflow Generator
|
| 3 |
emoji: π
|
| 4 |
+
colorFrom: blue
|
| 5 |
+
colorTo: purple
|
| 6 |
sdk: gradio
|
| 7 |
+
sdk_version: 4.44.0
|
| 8 |
app_file: app.py
|
| 9 |
+
pinned: true
|
| 10 |
license: apache-2.0
|
| 11 |
---
|
| 12 |
|
| 13 |
+
# π n8n Workflow Generator
|
| 14 |
|
| 15 |
+
Generate n8n workflows using natural language! Powered by fine-tuned **Qwen2.5-Coder-1.5B**.
|
| 16 |
|
| 17 |
+
## π― Features
|
| 18 |
|
| 19 |
+
- **Natural Language Input**: Describe your workflow in plain English
|
| 20 |
+
- **Instant Generation**: Get TypeScript DSL code in seconds
|
| 21 |
+
- **n8n JSON Export**: Ready to import into your n8n instance
|
| 22 |
+
- **Visual Preview**: See your workflow structure before using it
|
| 23 |
|
| 24 |
+
## π Performance
|
| 25 |
|
| 26 |
+
- **Overall Accuracy:** 91.8%
|
| 27 |
+
- **Test Cases:** 24 comprehensive workflow patterns
|
| 28 |
+
- **Grade A:** 83% of test cases
|
|
|
|
|
|
|
| 29 |
|
| 30 |
+
## π How to Use
|
| 31 |
|
| 32 |
+
1. Enter your workflow description (e.g., "Create a webhook that sends data to Slack")
|
| 33 |
+
2. Click "Generate Workflow"
|
| 34 |
+
3. Copy the generated code or n8n JSON
|
| 35 |
+
4. Import into your n8n instance
|
|
|
|
|
|
|
|
|
|
|
|
|
| 36 |
|
| 37 |
+
## π Example Prompts
|
|
|
|
|
|
|
| 38 |
|
| 39 |
- "Create a webhook that sends data to Slack"
|
| 40 |
+
- "Build a workflow that fetches GitHub issues and sends daily summary email"
|
| 41 |
+
- "Webhook receives order, if amount > $1000 send to priority queue"
|
| 42 |
+
- "Schedule that runs every Monday, fetches API data, updates Google Sheets"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 43 |
|
| 44 |
+
## π Links
|
| 45 |
|
| 46 |
+
- [Model Card](https://huggingface.co/Nishan30/n8n-workflow-generator)
|
| 47 |
+
- [GitHub Repository](https://github.com/Nishan30)
|
| 48 |
|
| 49 |
## π Acknowledgments
|
| 50 |
|
| 51 |
Built with:
|
| 52 |
- [Qwen2.5-Coder](https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B-Instruct)
|
| 53 |
+
- [Gradio](https://gradio.app/)
|
| 54 |
- [Hugging Face Transformers](https://github.com/huggingface/transformers)
|
|
|
|
|
|