---
title: moon-iclr2027-data
canonical_url: "https://www.modelscope.cn/datasets/jmzhou/moon-iclr2027-data"
md_url: "https://www.modelscope.cn/datasets/jmzhou/moon-iclr2027-data.md"
repository: jmzhou/moon-iclr2027-data
last_updated: 2026-10-30
license: "Apache License 2.0"
storage_size: "336 GB"
downloads: 110
stars: 0
---

# moon-iclr2027-data

> moon-iclr2027-data - jmzhou 在 ModelScope 开源的数据集。Learning When to Forget and How to Move On in Vision-and-Language Navigation

jmzhou/moon-iclr2027-data 是 ModelScope 魔搭社区上的数据集，存储大小 336 GB，采用 Apache License 2.0 许可。

- **Repository**: jmzhou/moon-iclr2027-data
- **License**: Apache License 2.0
- **Storage size**: 336 GB
- **Downloads**: 110
- **Stars**: 0
- **Last updated**: 2026-10-30

Source: https://www.modelscope.cn/datasets/jmzhou/moon-iclr2027-data

---

# MOON Data

**Learning When to Forget and How to Move On in Vision-and-Language Navigation**

This repository contains navigation episodes, scene assets, rollouts, and experiment files used in MOON.

## 📦 Data Files

| Directory | Contents |
|---|---|
| `datasets/` | R2R-CE, RxR-CE, EnvDrop, and ScaleVLN episode files, including English RxR splits. |
| `scenes/` | Matterport3D and HM3D scene assets. |
| `rollouts_stage2_v2/` | Navigation rollouts collected with the frozen base policy. |
| `expert_rollouts/` | Shortest-path expert trajectories and observation frames. |
| `core/`, `method1/`, `ablation_runs/` | Evaluation, analysis, and ablation records, along with judge examples and recovery starting states. |
| `envs/` | Packed Conda environments and dependency specifications. |
| `hope/`, `docs/`, `notes/` | Experiment scripts and supporting documentation. |

Files are distributed as archives, with numbered parts for larger packages. Use the included `SHA256SUMS` files to verify downloads.

## 🔗 Code

Source code, data preparation, and inference and evaluation instructions: [AutoLab-SAI-SJTU/MOON](https://github.com/AutoLab-SAI-SJTU/MOON).