---
title: Giga-World-1
canonical_url: "https://www.modelscope.cn/models/GigaAI/Giga-World-1"
md_url: "https://www.modelscope.cn/models/GigaAI/Giga-World-1.md"
repository: GigaAI/Giga-World-1
chinese_name: Giga-World-1
last_updated: 2026-07-30
license: apache-2.0
pipeline_tag: robotics
tasks:
- robotics
parameters: 47.4B
tensor_type:
- BF16
- F32
library_name:
- safetensors
- diffusers
- pytorch
frameworks:
- Pytorch
downloads: 758
stars: 0
---
# Giga-World-1
> Giga-World-1 - GigaAI 在 ModelScope 开源的模型。This repository contains checkpoints for GigaWorld-1, a world model specially optimized for robot policy evaluation, introduced in the paper GigaWorld-1: A Roadmap to Build World Models for Robot Policy Evaluation.
GigaAI/Giga-World-1 是 ModelScope 魔搭社区上的 47.4B 参数robotics模型，采用 apache-2.0 许可。
- **Repository**: GigaAI/Giga-World-1
- **License**: apache-2.0
- **Tasks**: robotics
- **Parameters**: 47.4B
- **Downloads**: 758
- **Stars**: 0
- **Last updated**: 2026-07-30
Source: https://www.modelscope.cn/models/GigaAI/Giga-World-1
---
# Giga-World-1
This repository contains checkpoints for **GigaWorld-1**, a world model specially optimized for robot policy evaluation, introduced in the paper [GigaWorld-1: A Roadmap to Build World Models for Robot Policy Evaluation](https://huggingface.co/papers/2607.02642).
**GigaWorld-1** provides open-source training, inference, data processing, checkpoint conversion, and LoRA merge workflows for robot world models.
- **Paper:** [GigaWorld-1: A Roadmap to Build World Models for Robot Policy Evaluation](https://huggingface.co/papers/2607.02642)
- **Project Page:** [Project Website](https://open-gigaai.github.io/giga-world-1/)
- **Repository:** [GitHub](https://github.com/open-gigaai/giga-world-1)
## Directory Structure
```text
model/
├── README.md
├── assets/
│ └── main_page.png
├── before_stage1/ # Base checkpoints before stage-1 training
│ ├── Wan2p1_1p3B-FunContro-GigaRobo-alpha-diffusers/
│ ├── Wan2p1_1p3B-FunControl-diffusers/
│ └── Wan2p2_5B-FunControl-diffusers/
├── stage1/ # Stage-1 fine-tuned checkpoints
│ ├── nano/ # small (1.3B) variant
│ └── pro/ # large (5B) variant
└── stage2_distill/ # Stage-2 distilled checkpoints
└── ...
```
### Training pipeline overview
### Model Overview
| Stage | Name | Path | Notes |
| --------------- | ----------------------------------------------- | ---------------------------------------------------- | -------------------------------------- |
| Open Source | WAN 2.1 1.3B FunControl | [Wan2.1-Fun-1.3B-Control](https://www.bing.com/search?q=WAN+2.1+1.3B+FunControl&form=ANNTH1&refig=6a436fa312dd427c9dfef065aca873ed&pc=CNNDDB)  | Open-source model. |
| Open Source | WAN 2.2 5B FunControl | [Wan2.2-Fun-5B-Control](https://huggingface.co/alibaba-pai/Wan2.2-Fun-5B-Control)  | Open-source model. |
| Before Stage 1 | GigaRobo Alpha Diffusers | `before_stage1/Wan2p1_1p3B-FunContro-GigaRobo-alpha-diffusers/` | Pretrained on Giga dataset, then converted to Diffusers. |
| Before Stage 1 | WAN 2.1 1.3B Diffusers | `before_stage1/Wan2p1_1p3B-FunControl-diffusers/` | Vanilla Diffusers-converted checkpoint.|
| Before Stage 1 | WAN 2.2 5B Diffusers | `before_stage1/Wan2p2_5B-FunControl-diffusers/` | Vanilla Diffusers-converted checkpoint.|
| Stage 1 | Nano (1.3B) | `stage1/nano/` | Stage-1 fine-tuned from the 1.3B branches. |
| Stage 1 | Pro (5B) | `stage1/pro/` | Stage-1 fine-tuned from the 5B branch. |
| Stage 2 | Nano Distill | 🚧 Coming soon | 🚧 Coming soon. |
| Stage 2 | Pro Distill | 🚧 Coming soon | 🚧 Coming soon. |
### Stage-1 checkpoint structure
Each Stage-1 variant (`nano` / `pro`) contains two released artifacts: a full Diffusers-format checkpoint and a scene LoRA checkpoint.
```text
stage1/{nano,pro}/
├── Giga-World-1-*-stage1_final-diffusers/ # full Diffusers checkpoint
│ ├── model_index.json # Diffusers pipeline index
│ ├── transformer/ # DiT / video transformer weights
│ ├── vae/ # VAE weights
│ ├── text_encoder/ # text encoder weights
│ ├── tokenizer/ # tokenizer files
│ ├── scheduler/ # scheduler config
│ ├── image_encoder/ # image encoder weights
│ └── image_processor/ # image preprocessing config
└── Giga-World-1-*-stage1_scene_lora/ # scene LoRA checkpoint
├── pytorch_lora_weights.safetensors # LoRA weights for inference
├── transformer_full/ # full transformer export
├── transformer_partial.pth # partial transformer checkpoint
├── pytorch_model/ # training checkpoint shards
├── distributed_checkpoint/ # distributed training checkpoint
├── scheduler.bin # training scheduler state
├── latest # latest checkpoint pointer
├── zero_to_fp32.py # ZeRO checkpoint conversion script
└── random_states_*.pkl # training random states
```
## Quick Start
### Hugging Face Repository
```text
https://huggingface.co/GigaAI-Research/Giga-World-1
```
### SDK Download
```bash
# Install Hugging Face Hub
pip install huggingface_hub
```
```python
# Download the model snapshot via Hugging Face Hub
from huggingface_hub import snapshot_download
model_dir = snapshot_download(repo_id='GigaAI-Research/Giga-World-1')
```
### Git Download
```bash
git lfs install
git clone https://huggingface.co/GigaAI-Research/Giga-World-1
```
## Acknowledgements
We sincerely thank the open-source community and the projects that make this work possible.
Thanks also to many other open-source contributors for their tools, models, and community support.
## Citation
Please cite the paper if you find it helpful:
```bibtex
@article{gigaworld2025,
title = {GigaWorld-1: A Roadmap to World Models for Robot Policy Evaluation},
author = {{GigaAI}},
journal = {arXiv preprint},
year = {2025},
eprint = {2607.02642},
archivePrefix = {arXiv},
primaryClass = {cs.CV}
}
```
## License
This model is released under the Apache License 2.0 unless otherwise specified.