Update README.md

README.md

@@ -1,12 +1,67 @@
 ---
 license: apache-2.0
+task_categories:
+- text-generation
 language:
 - en
 tags:
 - canis-teach
 - education
+- tutoring
 - math
-- dataset
+- synthetic
+- conversational
 size_categories:
 - 1K<n<10K
----
+---
+
+# Canis.teach Math Dataset
+
+Simple synthetic dataset for training Math tutoring models.
+
+- **Project**: Canis.teach - Learning that fits.
+- **Subject**: Math
+- **Generated with**: [Canis.lab](https://github.com/crasyK/Canis.lab)
+- **Format**: Simple ID:content pairs
+
+## Dataset Structure
+
+```json
+{
+  "id": "unique_identifier",
+  "content": "tutoring conversation text"
+}
+```
+
+This dataset contains educational conversations focused on Math topics, designed to teach effective tutoring behavior rather than just providing direct answers.
+
+## Usage
+
+```python
+from datasets import load_dataset
+
+dataset = load_dataset("CanisAI/teach-math-dataset")
+```
+
+## Content
+
+The conversations emphasize:
+- Step-by-step explanations
+- Pedagogical guidance
+- Encouraging learning through hints
+- Subject-appropriate teaching methods
+
+## License
+
+Apache 2.0
+
+## Related Models
+
+| Repository | Description |
+|------------|-------------|
+| `CanisAI/teach-math-qwen3-4b-2507-r1` | LoRA adapters |
+| `CanisAI/teach-math-qwen3-4b-2507-r1-merged` | Merged model |
+| `CanisAI/teach-math-qwen3-4b-2507-r1-gguf` | GGUF quantized |
+
+---
+**Canis.teach** - Learning that fits.