datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
gemma3n-slicing-configsThis repository contains configurations to slice Gemma 3n E4B, which is enabled thanks to it being a MatFormer.
The E4B model can be sliced into small models, trading off quality and latency/compute requirements.
We recommend exploring the [MatFormer Lab](TODO: add link) to getting started with slicing Gemma 3n E4B yourself.
For each configuration, we calculate the MMLU accuracy.
Although these are not the only configurations possible, they are optimal configurations
identified by calculating… See the full description on the dataset page: https://huggingface.co/datasets/google/gemma3n-slicing-configs.KYS-Configs
KYS-Configs
Everything needed to re-run Know Your Sources: Data Selection Matters when Rewriting for
Data-Constrained Pretraining end to end: the rewriting prompts, the vLLM generation settings, the
Nanotron training configs, the shared initial weights, and the full evaluation harness with all raw
results.
Layout
prompts/ the rewriting, annotation and judging prompts
vllm/ the rewriting workers and their SLURM launchers
nanotron/ the 18 training… See the full description on the dataset page: https://huggingface.co/datasets/blab-jhu/KYS-Configs.