README.md

---
license: mit
library_name: pytorch
pipeline_tag: robotics
base_model: InternRobotics/InternW0-Base
tags:
- robotics
- world-model
- robot-learning
- manipulation
- flow-matching
- video-action-model
- libero
- InternW0
language:
- en
---

# InternW0-Libero

[![Paper](https://img.shields.io/badge/Paper-arXiv%202609.27656-b31b1b)](https://arxiv.org/abs/2609.27656)
[![Project](https://img.shields.io/badge/Project-InternW0-2f80ed)](https://internrobotics.github.io/internw0)
[![Code](https://img.shields.io/badge/Code-GitHub-181717?logo=github)](https://github.com/InternRobotics/InternW0)
[![Base model](https://img.shields.io/badge/Base-InternW0--Base-6b7280)](https://huggingface.co/InternRobotics/InternW0-Base)
[![Downloads last month](https://img.shields.io/badge/dynamic/json?url=https%3A%2F%2Fhuggingface.co%2Fapi%2Fmodels%2FInternRobotics%2FInternW0-Libero&query=%24.downloads&label=downloads%20last%20month&logo=huggingface)](https://huggingface.co/InternRobotics/InternW0-Libero)

## Model description

**InternW0-Libero** is the task-posttrained InternW0 checkpoint for LIBERO. It
is derived from the heterogeneous
[InternW0-Base](https://huggingface.co/InternRobotics/InternW0-Base) pretraining
checkpoint and posttrained with the R13 Goal+Long recipe. The selected step-600
checkpoint is evaluated on all four official LIBERO suites.

InternW0 couples a slow Wan2.2-TI2V-5B VideoDiT planner with a fast ActionDiT
executor. Observation-conditioned chunk K/V routing adapts cached future-video
context to newly observed states, so the model can reuse long-horizon
predictions while generating responsive action chunks.

This is a **native `w0` checkpoint**, not a standalone Transformers model. It is
intended for LIBERO evaluation or compatible continued posttraining.

## Files and integrity

| File | Purpose | Size | SHA-256 |
|---|---|---:|---|
| `InternW0-Libero.pt` | Canonical model-only native checkpoint | 14,705,454,471 bytes | `5f967259b1a8dfce90c871212630d3bc1945ce6ceb03dc96aae6c148891935c7` |
| `pytorch_model.bin` | Zero-copy LFS alias used for PyTorch download statistics | Same LFS object | `5f967259b1a8dfce90c871212630d3bc1945ce6ceb03dc96aae6c148891935c7` |
| `config.json` | Machine-readable model/download-counting manifest | 1,744 bytes | `6b60594bcba8a6b434ab7d2140c3751c8d2418b537bd8011b4e99a522ed718a0` |
| `checkpoint_metadata.json` | Embedded behavior/training/statistics/provenance export | 16,024 bytes | `8fc7772aceeb20754c8c563b5651e8043886b3d6fbfc05e7a415a0cffac77e19` |
| `normalization.json` | Matching LIBERO EEF10 action/proprio statistics | 3,543 bytes | `9f63fc9ad9208515358ae560b4b9eec192b7f99f5ce815bf049bad3b269b0301` |
| `posttrain_config.yaml` | Portable public-code posttraining config | 1,509 bytes | `99d52eebf290f3270f5ddb769349f9e7c25e006161982efa9eeca9b5525b61f2` |
| `evaluation_config.yaml` | Portable public-code evaluation config | 917 bytes | `0a8e957bff862ec98de9c8c890f580ed40e4304b45d5bf81b60c0e1b614d3e5c` |
| `source_training_config.yaml` | Historical resolved R13 training config | 6,276 bytes | `a9209138c249fd07bed47d6d95096d9b0ec14606ff771ad24d52b1539713835b` |
| `source_evaluation_config.yaml` | Historical resolved evaluation config | 7,372 bytes | `d70258ce4d10bf15e1957bb45da0e88b229dfe9cab1f919c0547b996cda9d99e` |
| `evaluation_summary.json` | Complete 2,000-rollout four-suite result | 9,207 bytes | `76824bd0b69308bbf7db73db1ab77f679939d873cba07ae0c2b5488da2fc1d8d` |
| `evaluation_meta.json` | Evaluation launch metadata | 859 bytes | `c9566017ef589087df53abe04725143b1d68e9ebd05e5b78743c55b17b7e3af7` |
| `PROVENANCE.md` | Selection, lineage, training, and evaluation provenance | 2,799 bytes | `65e5dba7faf5b18203f9dffeb4acaff935893a275da5e0f386bfd4486ce697fc` |
| `LICENSE` | MIT license | 1,205 bytes | `b1d3d0160bdd9748bfc4e8acb9ab8fbce7ab725c827fd6fc673261f57aae0306` |

`pytorch_model.bin` is a second Git-LFS path to the exact same object as
`InternW0-Libero.pt`. It does not duplicate the 14.7 GB payload.

## Checkpoint summary

- Selected step: `600`
- Serialization dtype: `bfloat16`
- Model-only checkpoint: yes
- Optimizer state: excluded
- Tensors: 1,977
- Parameter elements: 7,352,528,643
- Source heterogeneous pretraining step: `30000`
- Target domain: `libero/franka`
- Target soft prompts: K=8
- Canonical action/proprio dimensionality: 37
- Native EEF10 slots: 10
- Action horizon: 64
- Action chunk size: 16
- Video plan horizon: 96 action ticks
- Video/action stride: 8
- Maximum action offset: 15
- History frames: 0
- Action semantics: native delta EEF10, row-aligned

The checkpoint embeds a schema-v4 `behavior_config`, `training_recipe`,
`posttrain_statistics`, and `posttrain_provenance`. The public loader performs a
semantic contract check before loading weights.

## LIBERO training

The R13 recipe uses:

- posttraining suites `libero_goal` and `libero_10`;
- random-frame sampling with 100,000 samples per epoch;
- 10 epochs;
- batch size 48 per process;
- learning rate 5e-5 with cosine decay;
- seed 42;
- native 20 Hz EEF10 actions and proprio;
- action offset sampled in `[0, 15]`;
- action horizon 64 with four 16-action chunks;
- K=8 target soft prompts.

The matching normalization statistics are in `normalization.json`. The R13 run
initialized from a preceding LIBERO checkpoint at step 2,640 and selected step
600 over the final step 660 based on the complete four-suite evaluation. See
`PROVENANCE.md` for the full lineage.

## Evaluation result

The selected checkpoint was evaluated on all 40 tasks in the four official
LIBERO suites, with 50 trials per task:

| Spatial | Object | Goal | Long (`libero_10`) | Average | Successes |
|---:|---:|---:|---:|---:|---:|
| **99.4%** | **99.4%** | **98.6%** | **97.0%** | **98.6%** | **1,972 / 2,000** |

The complete per-task record is included in `evaluation_summary.json`. The
evaluation uses environment/init seed 42, an unset policy diffusion seed, 30
settling steps, execution of all 16 actions per chunk, 10 inference steps,
runtime gripper binarization, lossless PNG transport, 400-step limits for
Spatial/Object/Goal, a 700-step limit for `libero_10`, and MuJoCo 3.3.2.

## Required external assets

This checkpoint is not self-contained. Users must separately obtain and comply
with the licenses for:

```bash
hf download Wan-AI/Wan2.2-TI2V-5B --local-dir assets/wan22
hf download facebook/dinov3-vitb16-pretrain-lvd1689m \
  --local-dir assets/dinov3-vitb16-pretrain-lvd1689m
```

A LIBERO checkout, the matching LIBERO LeRobot conversion, and its UMT5 text
cache are also required. The repository does not bundle datasets, simulation
code, Wan2.2 files, or DINOv3 weights.

## Quick start

### 1. Install the public codebase

```bash
git clone https://github.com/InternRobotics/InternW0.git
cd InternW0

conda create -n w0 python=3.10 -y
conda activate w0
pip install -U pip setuptools wheel
pip install torch==2.7.1 torchvision==0.22.1 \
  --extra-index-url https://download.pytorch.org/whl/cu128
pip install -e .
```

### 2. Download this checkpoint

```bash
hf download InternRobotics/InternW0-Libero \
  config.json pytorch_model.bin checkpoint_metadata.json \
  normalization.json posttrain_config.yaml evaluation_config.yaml \
  source_training_config.yaml source_evaluation_config.yaml \
  evaluation_summary.json evaluation_meta.json PROVENANCE.md LICENSE \
  --local-dir assets/libero

mv assets/libero/pytorch_model.bin \
   assets/libero/InternW0-Libero.pt
```

### 3. Run LIBERO evaluation

```bash
export CKPT="$PWD/assets/libero/InternW0-Libero.pt"
export W0_WAN_MODEL_DIR="$PWD/assets/wan22"
export W0_DINO_WEIGHTS="$PWD/assets/dinov3-vitb16-pretrain-lvd1689m"
export W0_LIBERO_STATS="$PWD/assets/libero/normalization.json"
export W0_LIBERO_TEXT_CACHE="$PWD/assets/libero/text_embeds_cache"
export LIBERO_PATH="$PWD/external/LIBERO"

SUITE=libero_goal TASK_ID=0 NUM_TRIALS=50 \
bash scripts/run_libero_eval.sh
```

For compatible continued posttraining:

```bash
bash scripts/train_zero1.sh 8 --config-name posttrain_libero
```

## Download statistics

Hugging Face uses library-specific query files for model download statistics.
For the `pytorch` library, the Hub counts HTTP `GET` and `HEAD` requests to
`pytorch_model.bin`. This repository provides that path as a zero-copy alias of
the canonical checkpoint. The badge above reads the public Hub API `downloads`
field, corresponding to downloads in the last 30 days. Hub analytics are
aggregated and may lag briefly.

## Intended uses

- LIBERO reproduction and extension.
- Research on world-action models, context routing, and cross-embodiment robot
  learning.
- Compatible continued posttraining with the public `w0` codebase.

## Limitations

- This is a simulation-domain research checkpoint, not a deployed robot
  controller.
- It requires the public InternW0 codebase, external Wan2.2 and DINOv3 assets,
  LIBERO data, text caches, and substantial compute.
- It is not directly compatible with `transformers.AutoModel`.
- Real-robot use requires independent safety analysis, constrained workspaces,
  and human oversight.

## License

The weights and metadata in this repository are released under the
[MIT License](./LICENSE). Third-party assets obtained separately remain subject
to their original licenses.

## Citation

If you use InternW0, please cite:

```bibtex
@article{internw0,
  title = {InternW0: A Foundational Physical World Model for Efficient Real-World Interactions},
  author = {Jisong Cai and Yao Mu and Ganlin Yang and Zhe Cao and Zhangzheng Tu and Xing Gao and Kailin Li and Xinyu Zhan and Lixin Yang and Yangkun Zhu and Haoxiang Ma and Ming Zhou and Qiaojun Yu and Yufei Xue and Liqun He and Yifei Yao and Yifan Zhu and Long Ling and Bingqi Jiang and Haoyu Guo and Xueyue Zhu and Bowen Zhou and Bin Zhao and Tianfan Xue and Chunhua Shen and Weinan Zhang},
  journal = {arXiv preprint arXiv:2609.27656},
  year = {2026}
}
```