This model has been finetuned on code extracted from over 200 of my personal GitHub repositories, building upon the deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B base model. It’s optimized for both text-to-text and coding applications.
Training Details:
Estimated Tokens Trained On: ~20 million tokens
Dataset generation was achieved using the GIT_HISTORY tool.
