Datasets:

---

llama-duo

/

gemma2b-summarize-eval-by-claude3sonnet

like 0

Follow

llama-duo 17

The default config contains 39 while it should generally contain 3 splits maximum (train/validation/test). If the splits gemma2b_summarize_gemini1_5flash_1k_by_claude3sonnet, gemma2b_summarize_gemini1_5flash_2k_by_claude3sonnet, gemma2b_summarize_gemini1_5flash_4k_by_claude3sonnet, gemma2b_summarize_gemini1_5flash_8k_by_claude3sonnet, gemma2b_summarize_gemini1_5flash_16k_by_claude3sonnet... are not used to differentiate between training and evaluation, please consider defining configs of this dataset instead. You can find how to define configs instead of splits here: https://huggingface.co/docs/hub/datasets-data-files-configuration