Dataset automatically created during the evaluation run of model Qwen/Qwen1.5-MoE-A2.7B.
The dataset is composed of 136 configuration, each one coresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
An additional… See the full description on the dataset page:
https://huggingface.co/datasets/OALL/details_Qwen__Qwen1.5-MoE-A2.7B.
