README.md

Fine-tuned DeepSeek-R1-Distill-Qwen-14B

This space hosts a fine-tuned version of the unsloth/DeepSeek-R1-Distill-Qwen-14B-bnb-4bit model.

Model Details

• Base Model: unsloth/DeepSeek-R1-Distill-Qwen-14B-bnb-4bit
• Fine-tuned on: phi4-cognitive-dataset
• Quantization: Already 4-bit quantized (no additional quantization applied)

Current Status

This space is currently being prepared. The fine-tuned model will be available soon.

Usage

Once deployed, you can interact with the model through the Gradio interface or via API.

Training Process

The model is being fine-tuned with the following specifications:

• Training dataset processed in ascending order by prompt_number
• Custom training parameters optimized for the L40S GPU
• Mixed precision training for optimal performance

Contact

For questions or issues, please reach out through the Hugging Face community.