--- license: llama3.2 datasets: - openai/gsm8k base_model: - meta-llama/Llama-3.2-1B --- llama 3.2-1B trained on 800 steps of gsm8k with Unsloth.