🚀 UGround-V1-2B （基于Qwen2-VL）

UGround是一个强大的GUI视觉定位模型，采用简单的方法进行训练。更多详情请查看我们的主页和论文。这项工作是OSUNLP和Orby AI合作完成的。

• 主页：https://osu-nlp-group.github.io/UGround/
• 仓库：https://github.com/OSU-NLP-Group/UGround
• 论文（ICLR'25口头报告）：https://arxiv.org/abs/2410.05243
• 演示：https://huggingface.co/spaces/orby-osu/UGround
• 联系人：苟博宇

✨ 主要特性

• UGround是一个强大的GUI视觉定位模型，使用简单的训练方法。
• 与OSUNLP和Orby AI合作完成。
• 提供了模型、主页、仓库、论文、演示等相关资源。

📦 安装指南

文档中未提及安装步骤，故跳过此章节。

💻 使用示例

推理示例

vLLM服务器

vllm serve osunlp/UGround-V1-7B  --api-key token-abc123 --dtype float16

或者

python -m vllm.entrypoints.openai.api_server --served-model-name osunlp/UGround-V1-7B --model osunlp/UGround-V1-7B --dtype float16 

你可以在Qwen2-VL的官方仓库中找到更多关于训练和推理的说明。

视觉定位提示

def format_openai_template(description: str, base64_image):
    return [
        {
            "role": "user",
            "content": [
                {
                    "type": "image_url",
                    "image_url": {"url": f"data:image/jpeg;base64,{base64_image}"},
                },
                {
                    "type": "text",
                    "text": f"""
  Your task is to help the user identify the precise coordinates (x, y) of a specific area/element/object on the screen based on a description.

  - Your response should aim to point to the center or a representative point within the described area/element/object as accurately as possible.
  - If the description is unclear or ambiguous, infer the most relevant area or element based on its likely context or purpose.
  - Your answer should be a single string (x, y) corresponding to the point of the interest.

  Description: {description}

  Answer:"""
                },
            ],
        },
    ]

messages = format_openai_template(description, base64_image)

completion = await client.chat.completions.create(
    model=args.model_path,
    messages=messages,
    temperature=0  # REMEMBER to set temperature to ZERO!
# REMEMBER to set temperature to ZERO!
# REMEMBER to set temperature to ZERO!
)

# The output will be in the range of [0,1000), which is compatible with the original Qwen2-VL
# So the actual coordinates should be (x/1000*width, y/1000*height)

Qwen2-VL-2B-Instruct使用示例

from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor
from qwen_vl_utils import process_vision_info

# default: Load the model on the available device(s)
model = Qwen2VLForConditionalGeneration.from_pretrained(
    "Qwen/Qwen2-VL-2B-Instruct", torch_dtype="auto", device_map="auto"
)

# We recommend enabling flash_attention_2 for better acceleration and memory saving, especially in multi-image and video scenarios.
# model = Qwen2VLForConditionalGeneration.from_pretrained(
#     "Qwen/Qwen2-VL-2B-Instruct",
#     torch_dtype=torch.bfloat16,
#     attn_implementation="flash_attention_2",
#     device_map="auto",
# )

# default processer
processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct")

# The default range for the number of visual tokens per image in the model is 4-16384. You can set min_pixels and max_pixels according to your needs, such as a token count range of 256-1280, to balance speed and memory usage.
# min_pixels = 256*28*28
# max_pixels = 1280*28*28
# processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct", min_pixels=min_pixels, max_pixels=max_pixels)

messages = [
    {
        "role": "user",
        "content": [
            {
                "type": "image",
                "image": "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg",
            },
            {"type": "text", "text": "Describe this image."},
        ],
    }
]

# Preparation for inference
text = processor.apply_chat_template(
    messages, tokenize=False, add_generation_prompt=True
)
image_inputs, video_inputs = process_vision_info(messages)
inputs = processor(
    text=[text],
    images=image_inputs,
    videos=video_inputs,
    padding=True,
    return_tensors="pt",
)
inputs = inputs.to("cuda")

# Inference: Generation of the output
generated_ids = model.generate(**inputs, max_new_tokens=128)
generated_ids_trimmed = [
    out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
]
output_text = processor.batch_decode(
    generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
)
print(output_text)

from PIL import Image
import requests
import torch
from torchvision import io
from typing import Dict
from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor

# Load the model in half-precision on the available device(s)
model = Qwen2VLForConditionalGeneration.from_pretrained(
    "Qwen/Qwen2-VL-2B-Instruct", torch_dtype="auto", device_map="auto"
)
processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct")

# Image
url = "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg"
image = Image.open(requests.get(url, stream=True).raw)

conversation = [
    {
        "role": "user",
        "content": [
            {
                "type": "image",
            },
            {"type": "text", "text": "Describe this image."},
        ],
    }
]


# Preprocess the inputs
text_prompt = processor.apply_chat_template(conversation, add_generation_prompt=True)
# Excepted output: '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n<|im_start|>user\n<|vision_start|><|image_pad|><|vision_end|>Describe this image.<|im_end|>\n<|im_start|>assistant\n'

inputs = processor(
    text=[text_prompt], images=[image], padding=True, return_tensors="pt"
)
inputs = inputs.to("cuda")

# Inference: Generation

📚 详细文档

模型信息

• Model-V1：
  • 初始UGround
  • UGround-V1-2B (Qwen2-VL)
  • UGround-V1-7B (Qwen2-VL)
  • UGround-V1-72B (Qwen2-VL)
  • 训练数据

发布计划

• 模型权重：
  • [x] 初始版本（论文中使用的版本）
  • [x] 基于Qwen2-VL的V1版本（2B、7B、72B）
• 代码：
  • [x] UGround的推理代码（初始版本和基于Qwen2-VL的版本）
  • [x] 离线实验（代码、结果和有用资源）
    • [x] ScreenSpot
    • [x] Multimodal-Mind2Web
    • [x] OmniAct
    • [x] Android Control
  • [x] 在线实验
    • [x] Mind2Web-Live-SeeAct-V
    • [x] AndroidWorld-SeeAct-V
  • [ ] 数据合成管道（即将推出）
• 训练数据（V1）：https://huggingface.co/datasets/osunlp/UGround-V1-Data
• 在线演示（HF Spaces）

主要结果

GUI视觉定位：ScreenSpot（标准设置）

GUI视觉定位：ScreenSpot（代理设置）

Qwen2-VL-2B-Instruct介绍

新特性

• 对各种分辨率和比例图像的最优理解：Qwen2-VL在视觉理解基准测试中取得了最先进的性能，包括MathVista、DocVQA、RealWorldQA、MTVQA等。
• 理解20分钟以上的视频：Qwen2-VL可以理解超过20分钟的视频，用于高质量的基于视频的问答、对话、内容创作等。
• 可操作移动设备、机器人等的代理：具有复杂推理和决策能力，Qwen2-VL可以与手机、机器人等设备集成，根据视觉环境和文本指令进行自动操作。
• 多语言支持：为了服务全球用户，除了英语和中文，Qwen2-VL现在支持理解图像中不同语言的文本，包括大多数欧洲语言、日语、韩语、阿拉伯语、越南语等。

模型架构更新

• 朴素动态分辨率：与以前不同，Qwen2-VL可以处理任意图像分辨率，将其映射到动态数量的视觉标记，提供更像人类的视觉处理体验。

我们有三个参数分别为20亿、70亿和720亿的模型。此仓库包含经过指令微调的2B Qwen2-VL模型。更多信息，请访问我们的博客和GitHub。

评估

图像基准测试

视频基准测试

要求

Qwen2-VL的代码已包含在最新的Hugging face transformers中，我们建议你使用以下命令从源代码构建：

pip install git+https://github.com/huggingface/transformers

否则，你可能会遇到以下错误：

KeyError: 'qwen2_vl'

快速开始

我们提供了一个工具包，帮助你更方便地处理各种类型的视觉输入，包括base64编码、URL和交错的图像和视频。你可以使用以下命令安装：

pip install qwen-vl-utils

🔧 技术细节

文档中未提及技术细节相关内容，故跳过此章节。

📄 许可证

本项目采用Apache-2.0许可证。

📚 引用信息

如果你发现这项工作有用，请考虑引用我们的论文：

@article{gou2024uground,
        title={Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents},
        author={Boyu Gou and Ruohan Wang and Boyuan Zheng and Yanan Xie and Cheng Chang and Yiheng Shu and Huan Sun and Yu Su},
        journal={arXiv preprint arXiv:2410.05243},
        year={2024},
        url={https://arxiv.org/abs/2410.05243},
      }

@article{zheng2023seeact,
        title={GPT-4V(ision) is a Generalist Web Agent, if Grounded},
        author={Boyuan Zheng and Boyu Gou and Jihyung Kil and Huan Sun and Yu Su},
        journal={arXiv preprint arXiv:2401.01614},
        year={2024},
      }