Pre-trained agents for Hanabi

该数据集包含了一系列经过预训练的智能体，包括自我对弈智能体和针对零样本协同（ZSC）进行训练的智能体，其目的是为了评估多智能体强化学习（MARL）方法的适应性。数据集涵盖了多种算法，如IQL、VDN、OBL等，并记录了它们在多次适应性试验中的性能指标。该任务旨在合作多智能体环境中对适应性进行评估。

This dataset contains a series of pre-trained agents, including self-play agents and agents trained for zero-shot collaboration (ZSC), with the goal of evaluating the adaptability of multi-agent reinforcement learning (MARL) methods. The dataset covers multiple algorithms such as IQL, VDN, OBL, and others, and records their performance metrics across numerous adaptation trials. This task is designed to assess adaptability in cooperative multi-agent environments.