nanocore-base-v0.1 is a fully fine-tuned version of HuggingFaceTB/SmolLM2-135M.
Trained with Axolotl on a custom dataset of 1,000 written instruction examples.
Dataset covers topics like mathematics, science, logic, problem solving, and instruction following.
Intended uses are lightweight text generation, and local automation tasks.
Limitations: Due to its 135M parameter size, complex understanding may be limited.