SmolLM，这是一系列最先进的小型模型，具有 135M、360M 和 1.7B 参数，并在新的高质量数据集上进行了训练。它涵盖数据管理、模型评估和使用。

参考文献：
[1] SmolLM - blazingly fast and remarkably powerful：https://huggingface.co/blog/smollm
[2] https://huggingface.co/HuggingFaceTB/SmolLM-135M
[3] https://huggingface.co/HuggingFaceTB/SmolLM-360M
[4] https://huggingface.co/HuggingFaceTB/SmolLM-1.7B
[5] https://huggingface.co/datasets/HuggingFaceTB/smollm-corpus
[6] SmolLM models collection: https://huggingface.co/collections/HuggingFaceTB/smollm-models-6695016cad7167254ce15966
[7] SmolLM-Corpus dataset: https://huggingface.co/datasets/HuggingFaceTB/smollm-corpus
[8] WebGPU demo: https://huggingface.co/spaces/HuggingFaceTB/SmolLM-135M-Instruct-WebGPU and https://huggingface.co/spaces/HuggingFaceTB/SmolLM-360M-Instruct-WebGPU