🤏 Smol-Data Collection Tried and tested mixes for strong pretraining. Inspired by https://huggingface.co/blog/codelion/optimal-dataset-mixing • 14 items • Updated 11 days ago • 12
Claude 4.5 Opus Collection Distilled models and datasets for Claude 4.5 Opus. • 14 items • Updated 11 days ago • 29
PockEngine: Sparse and Efficient Fine-tuning in a Pocket Paper • 2310.17752 • Published Oct 26, 2023 • 15