Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
The first behavioral foundation model to control a virtual physics-based humanoid agent for a wide range of whole-body tasks.
| Date | Stars |
|---|---|
| 2026-07-31 | 778 |
| 2026-08-06 | 778 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# Meta Motivo
**[Meta, FAIR](https://ai.facebook.com/research/)**
# Overview
This repository provides a PyTorch implementation and pre-trained models for Meta Motivo. For details see the paper [Zero-Shot Whole-Body Humanoid Control via Behavioral Foundation Models](https://metamotivo.metademolab.com/).
### Features
- We provide [**6** pretrained FB-CPR](https://huggingface.co/collections/facebook/meta-motivo-6757761e8fd4a032466fd129) models for controlling the humanoid model defined in [HumEnv](https://github.com/facebookresearch/HumEnv/).
- **Fully reproducible** scripts for evaluating the model in HumEnv.
- **Fully reproducible** [FB-CPR training code in HumEnv](examples/fbcpr_train_humenv.py) for the full results in the paper, and [FB training code in DMC](examples/fb_train_dmc.py) for faster experimentation.
# Installation
The project is pip installable in your environment.
```
pip install "metamotivo[huggingface,humenv] @ git+https://github.com/facebookresearch/metamotivo.git"
```
It requires Python 3.10+. Optional dependencies include `humenv["bench"]` and `huggingface_hub` for testing/training and loading models from HuggingFace.
# Pretrained models
For reproducibility, we provide all the **5** models (**metamotivo-S-X**) we trained for producing the results in the [paper](https://openreview.net/forum?id=9sOR0nYLtz&referrer=%5BAuthor%20Console%5D(%2Fgroup%3Fid%3DICLR.cc%2F2025%2FConference%2FAuthors%23your-submissions)), where each model is trained using a different random seed. We also provide our largest and most performant model (**metamotivo-M-1**), which can also be interactively tested in our [demo](https://metamotivo.metademolab.com/).
| Model | # of params | Download |
| :--- | :---: | :---: |
| metamotivo-S-1 | 24.5M | [link](https://huggingface.co/facebook/metamotivo-S-1) |
| metamotivo-S-2 | 24.5M | [link](https://huggingface.co/facebook/metamotivo-S-2) |
| metamotivo-S-3 | 24.5M | [link](https://huggingface.co/facebook/metamotivo-S-3) |
| metamotivo-S-4 | 24.5M | [link](https://huggingface.co/facebook/metamotivo-S-4) |
| metamotivo-S-5 | 24.5M | [link](https://huggingface.co/facebook/metamotivo-S-5) |
| metamotivo-M-1 | 288M | [link](https://huggingface.co/facebook/metamotivo-M-1) |
# Quick start
Once the library is installed, you can easily create an FB-CPR agent and download a pre-trained model from the Hugging Face hub. Note that the model is an instance of `torch.nn.Module` and by default it is initialized in "inference" mode (no_grad and eval mode).
We provide some simple code snippets to demonstrate how to use the model below. For more detailed examples, see our tutorials on [interacting with the model](https://github.com/facebookresearch/metamotivo/blob/main/tutorial.ipynb), [running an evaluation](https://github.com/facebookresearch/metamotivo/blob/main/tutorial_benchmark.ipynb), and [training from scratch](https://github.com/facebookresearch/metamotivo/tree/main/examples).
### Download the pre-trained models
The following code snippet shows how to instantiate the model.
```python
from metamotivo.fb_cpr.huggingface import FBcprModel
model = FBcprModel.from_pretrained("facebook/metamotivo-S-1")
```
### Download the buffers
For each model we provide:
- The training buffer (that can be used for inference or offline training)
- A small reward inference buffer (that contains the minimum amount of information for doing reward inference)
```python
from huggingface_hub import hf_hub_download
import h5py
local_dir = "metamotivo-S-1-datasets"
dataset = "buffer_inference_500000.hdf5" # a smaller buffer that can be used for reward inference
# dataset = "buffer.hdf5" # the full training buffer of the model
buffer_path = hf_hub_download(
repo_id="facebook/metamotivo-S-1",
filename=f"data/{dataset}",
repo_type="model",
local_dir=local_dir,
Excerpt of 10,980 characters
Read on GitHub
5
Anssi · @facebookresearch · United Kingdom
2
1
1
1
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:b14631bbf1223504, desc:humanoid