README.md

---
license: apache-2.0
datasets:
- declare-lab/Emma-X-GCOT
metrics:
- accuracy
base_model:
- openvla/openvla-7b
pipeline_tag: image-text-to-text
---

<h1 align="center">✨ 
<br/>  
Meet Emma-X, an Embodied Multimodal Action Model 
<br/>
✨✨✨


</h1>

<div align="center">
  <img src="https://raw.githubusercontent.com/declare-lab/Emma-X/main/Emma-X.png" alt="Emma-X" width="300" />

<br/>

[![arXiv](https://img.shields.io/badge/arxiv-2412.11974-b31b1b)](https://arxiv.org/abs/2412.11974) [![Emma-X](https://img.shields.io/badge/Huggingface-Emma--X-brightgreen?style=flat&logo=huggingface&color=violet)](https://huggingface.co/declare-lab/Emma-X) [![Static Badge](https://img.shields.io/badge/Demos-declare--lab-brightred?style=flat)](https://declare-lab.github.io/Emma-X/)


</div>

## Model Overview

EMMA-X is an Embodied Multimodal Action (VLA) Model designed to bridge the gap between Visual-Language Models (VLMs) and robotic control tasks. EMMA-X generalizes effectively across diverse environments, objects, and instructions while excelling at long-horizon spatial reasoning and grounded task planning using a novel Trajectory Segmentation Strategy. It relies on --

- Hierarchical Embodiment Dataset: Emma-X is trained on a dataset derived from BridgeV2, containing 60,000 robot manipulation trajectories. Trained using a hierarchical dataset with visual grounded chain-of-thought reasoning, EMMA-X's output will include the following components:

- Grounded Chain-of-Thought Reasoning: Helps break down tasks into smaller, manageable subtasks, ensuring accurate task execution by mitigating hallucination in reasoning.

- Gripper Position Guidance: Affordance point inside the image.

- Look-Ahead Spatial Reasoning: Enables the model to plan actions while considering spatial guidance for effective planning, enhancing long-horizon task performance. 

It generates: 

- Action: Action policy in 7-dimensional vector to control the robot ([WidowX-6Dof](https://www.trossenrobotics.com/widowx-250)).

## Model Card
- **Developed by:** SUTD Declare Lab
- **Model type:** Vision-language-action (language, image => reasoning, robot actions)
- **Language(s) (NLP):** en
- **License:** Apache-2.0
- **Finetuned from:** [`openvla-7B`](https://huggingface.co/openvla/openvla-7b/)
- **Pretraining Dataset:** Augmented version of [Bridge V2](https://rail-berkeley.github.io/bridgedata/), for more info check our repository.
- **Repository:** [https://github.com/declare-lab/Emma-X/](https://github.com/declare-lab/Emma-X/)
- **Paper:** [Emma-X: An Embodied Multimodal Action Model with Grounded Chain of Thought and Look-ahead Spatial Reasoning](https://arxiv.org/pdf/2412.11974)
- **Project Page & Videos:** [https://declare-lab.github.io/Emma-X/](https://declare-lab.github.io/Emma-X/)

## Getting Started
```python
# Install minimal dependencies (`torch`, `transformers`, `timm`, `tokenizers`, ...)
# > pip install -r https://raw.githubusercontent.com/openvla/openvla/main/requirements-min.txt
from transformers import AutoModelForVision2Seq, AutoProcessor
from PIL import Image

import torch

# Load Emma-X
vla = AutoModelForVision2Seq.from_pretrained(
    "declare-lab/Emma-X",
    attn_implementation="flash_attention_2",  # [Optional] Requires `flash_attn`
    torch_dtype=torch.bfloat16, 
    low_cpu_mem_usage=True, 
    trust_remote_code=True
).to("cuda:0")

# Grab image input & format prompt of size 224x224
image: Image.Image = get_from_camera(...)
prompt = "In: What action should the robot take to achieve the instruction\nINSTRUCTION: \n{<Instruction here>}\n\nOut: "

# Predict Action (action is a 7 dimensional vector to control the robot)
action, grounded_reasoning = vla.generate_actions(
    image=image, prompt_text=prompt, type="act", do_sample=False,
    max_new_tokens=512, do_sample=False
)

print("Grounded Reasoning:", grounded_reasoning)
# Execute...
robot.act(action, ...)
```

## Citation
```
@article{sun2024emma,
  title={Emma-X: An Embodied Multimodal Action Model with Grounded Chain of Thought and Look-ahead Spatial Reasoning},
  author={Sun, Qi and Hong, Pengfei and Pala, Tej Deep and Toh, Vernon and Tan, U-Xuan and Ghosal, Deepanway and Poria, Soujanya},
  journal={arXiv preprint arXiv:2412.11974},
  year={2024}
}
```