Dominic/smollm135_fullprec_tinystories

SmolLM-135M model in full precision (BF16), fine-tuned on TinyStories. Trained for 12k steps on 200k train stories with eval on the published validation split (~6.3 perplexity).

See ./generate_tinystories_fullprec.py for simple demo. This model is only intended for generating toy story examples and comparing quantization techniques.