Add model card

README.md

@@ -0,0 +1,75 @@
+---
+license: apache-2.0
+base_model: unsloth/Qwen3-4B-Instruct-2507
+tags:
+- qwen3
+- performance-prediction
+- scope
+- unsloth
+- fine-tuned
+language:
+- en
+- zh
+---
+
+# SCOPE-Direct-v2
+
+This is a fine-tuned version of [unsloth/Qwen3-4B-Instruct-2507](https://huggingface.co/unsloth/Qwen3-4B-Instruct-2507) trained on the [SCOPE-sft-direct-data](https://huggingface.co/datasets/Cooolder/SCOPE-sft-direct-data) dataset.
+
+## Model Details
+
+- **Base Model**: Qwen3-4B-Instruct-2507
+- **Training Dataset**: SCOPE-sft-direct-data
+- **Training Method**: LoRA fine-tuning with Unsloth
+- **Merged**: Yes (16-bit merged model)
+
+## Training Configuration
+
+- Max Sequence Length: 4096
+- LoRA Rank: 32
+- Batch Size: 4
+- Gradient Accumulation Steps: 4
+- Learning Rate: 2e-5
+- Epochs: 1
+- Optimizer: adamw_8bit
+- LR Scheduler: cosine
+
+## Usage
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+model = AutoModelForCausalLM.from_pretrained(
+    "Cooolder/SCOPE-Direct-v2",
+    torch_dtype="auto",
+    device_map="auto"
+)
+tokenizer = AutoTokenizer.from_pretrained("Cooolder/SCOPE-Direct-v2")
+
+# Generate text
+messages = [
+    {"role": "user", "content": "Your prompt here"}
+]
+text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+inputs = tokenizer(text, return_tensors="pt").to(model.device)
+outputs = model.generate(**inputs, max_new_tokens=512)
+print(tokenizer.decode(outputs[0], skip_special_tokens=True))
+```
+
+## License
+
+This model inherits the license from its base model (Apache 2.0).
+
+## Citation
+
+If you use this model, please cite:
+
+```bibtex
+@misc{scope-direct-v2,
+  title={SCOPE-Direct-v2: Fine-tuned Qwen3-4B for Performance Prediction},
+  author={Your Name},
+  year={2025},
+  publisher={HuggingFace},
+  howpublished={\url{https://huggingface.co/Cooolder/SCOPE-Direct-v2}}
+}
+```