9 results

sycl

inall models datasets apps agents people projects

Datasets

All datasets matching “sycl”

BatsResearch /sycl SyCL Data This is the data used for experiments in Beyond Contrastive Learning: Synthetic Data Enables List-wise Training with Multiple Levels of Relevance paper. Synthetic Data The data is created from MS MARCO queries shared in mteb/msmarco repo. Directories llama33_70b, qwen25_72b, and qwen25_32b contain data generated with Llama 3.3 70B, Qwen2.5 72B, and Qwen2.5 32B, respectively. The data format in each subdirectory is the same as mteb/msmarco repo (trec20 is the… See the full description on the dataset page: https://huggingface.co/datasets/BatsResearch/sycl.tabular1M<n<10M2 likes110 downloads1y agoHugging Face