Update README.md

README.md

@@ -1,3 +1,59 @@
----
-license: cc-by-4.0
----
+---
+license: apache-2.0
+language:
+  - en
+tags:
+  - LLM
+  - classification
+  - instruction-tuned
+  - multi-label
+  - qwen
+datasets:
+  - custom
+pipeline_tag: text-classification
+---
+
+# BenchHub-Cat-7b
+
+**BenchHub-Cat-7b** is a 7B parameter instruction-tuned language model that performs structured classification of natural language queries into three dimensions:
+
+- `subject`: Topic domain of the query (e.g., law, health, travel)
+- `skill`: Type of skill or task (e.g., reasoning, explanation, comparison)
+- `target`: General or cultural-specific target audience
+
+It is based on the Qwen2.5-7B-Instruct architecture and trained on a mixture of synthetic and GPT-generated instruction data.
+
+## 🔧 Model Details
+
+- **Base Model**: Qwen2.5-7B-Instruct
+- **Task**: Structured triple-label classification
+- **Prompt Format**: Instruction-style with output structure
+- **Training Framework**: Axolotl + DeepSpeed ZeRO-3
+
+## 🧪 Training Configuration
+
+| Hyperparameter          | Value                |
+|--------------------------|----------------------|
+| Sequence Length          | 8192                 |
+| Learning Rate            | 2 × 10⁻⁵             |
+| Batch Size (Effective)   | 256                  |
+| Epochs                   | 3                    |
+| Scheduler                | Cosine Decay         |
+| Warmup Ratio             | 0.05                 |
+| Optimizer                | Method from [19]     |
+| Hardware                 | 4× A6000 48GB GPUs   |
+| Training Time            | ~5 hours per run     |
+
+## 🧠 Intended Use
+
+**Input**: Open-ended natural language queries  
+**Output**: Structured classification result with 3 fields
+
+### Example Categories:
+- `subject`: education, health, history, law, etc.
+- `skill`: reasoning, recall, summarization, etc.
+- `target`: general, cultural-specific
+
+### ✨ Example Prompt & Output
+
+#### 📝 Prompt