We build the SmolLM-Instruct (v0.2) models (135M, 360M and 1.7B) by doing SFT on a mix of these datasets:
- a dataset of 2k simple everyday conversations we generated by llama3.1-70B everyday-conversations-llama3.1-2k
- Magpie-Pro-300K-Filtered
- StarCoder2-Self-OSS-Instruct
- A small subset of OpenHermes-2.5
Follow the installation instructions in https://github.com/huggingface/alignment-handbook/tree/main?tab=readme-ov-file#installation-instructions
We train the models on 8 GPUs using the following command:
ACCELERATE_LOG_LEVEL=info accelerate launch --config_file recipes/accelerate_configs/deepspeed_zero3.yaml scripts/run_sft.py recipes/smollm/sft/config.yaml