forseasons/dabench-synthetic-pipeline-data

Name: dabench-synthetic-pipeline-data: dab_sandbox_extra_all_verl
Creator: forseasons

dabench-synthetic-pipeline-data: dab sandbox extra all verl: RL dataset on Hugging Face. Unified VERL-format DAB sandbox dataset assembled from the current extra DAB synthetic training packages. Code/pipeline source: AlkeraAI/DataAgent, branch dabench synthetic data v1.

forseasons/dabench-synthetic-pipeline-data on the Hugging Face Hub