qgallouedec/prj_gia_dataset_metaworld_stick_pull_v2_1111

---
library_name: gia
tags:
- deep-reinforcement-learning
- reinforcement-learning
- gia
- multi-task
- multi-modal
- imitation-learning
- offline-reinforcement-learning
---

An imitation learning environment for the stick-pull-v2 environment, sample for the policy stick-pull-v2

This environment was created as part of the Generally Intelligent Agents project gia: https://github.com/huggingface/gia

## Load dataset

First, clone it with

```sh
git clone https://huggingface.co/datasets/qgallouedec/prj_gia_dataset_metaworld_stick_pull_v2_1111
```

Then, load it with

```python
import numpy as np
dataset = np.load("prj_gia_dataset_metaworld_stick_pull_v2_1111/dataset.npy", allow_pickle=True).item()
print(dataset.keys()) # dict_keys(['observations', 'actions', 'dones', 'rewards'])
```

原始信息汇总

数据集概述

标签

• 深度强化学习
• 强化学习
• 多任务
• 多模态
• 模仿学习
• 离线强化学习

描述

这是一个用于stick-pull-v2环境的模仿学习环境，样本用于stick-pull-v2策略。

加载数据集

1. 克隆数据集： sh git clone https://huggingface.co/datasets/qgallouedec/prj_gia_dataset_metaworld_stick_pull_v2_1111

2. 加载数据集： python import numpy as np dataset = np.load("prj_gia_dataset_metaworld_stick_pull_v2_1111/dataset.npy", allow_pickle=True).item() print(dataset.keys()) # dict_keys([observations, actions, dones, rewards])