GTA1 32B

🚀 GTA1: State-of-the-Art GUI Grounding Models

This project shares state-of-the-art GUI grounding models trained using GRPO, which directly incentivizes actionable and grounded responses in reinforcement learning.

🚀 Quick Start

Reinforcement learning (RL) (e.g., GRPO) aids in grounding due to its inherent objective alignment—rewarding successful clicks—instead of promoting long textual Chain-of-Thought (CoT) reasoning. Unlike methods that rely heavily on verbose CoT reasoning, GRPO directly encourages actionable and grounded responses. Based on the findings from our blog, we present state-of-the-art GUI grounding models trained with GRPO.

✨ Features

• Objective Alignment: GRPO rewards successful clicks, directly incentivizing grounded responses.
• State-of-the-Art Performance: Achieves the best results among all open - source model families on challenging datasets.

📦 Installation

No installation steps are provided in the original document, so this section is skipped.

💻 Usage Examples

Basic Usage

from PIL import Image
from qwen_vl_utils import process_vision_info, smart_resize
from transformers import Qwen2_5_VLForConditionalGeneration, AutoProcessor
import torch
import re

SYSTEM_PROMPT = '''
You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. The image resolution is height {height} and width {width}. For elements with area, return the center point.

Output the coordinate pair exactly:
(x,y)
'''
SYSTEM_PROMPT=SYSTEM_PROMPT.strip()

# Function to extract coordinates from model output
def extract_coordinates(raw_string):
    try:
        matches = re.findall(r"\((-?\d*\.?\d+),\s*(-?\d*\.?\d+)\)", raw_string)
        return [tuple(map(int, match)) for match in matches][0]
    except:
        return 0,0

# Load model and processor
model_path = "HelloKKMe/GTA1-32B"
max_new_tokens = 32

model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
    model_path,
    torch_dtype=torch.bfloat16,
    attn_implementation="flash_attention_2",
    device_map="auto"
)
processor = AutoProcessor.from_pretrained(
    model_path,
    min_pixels=3136,
    max_pixels= 4096 * 2160
)

# Load and resize image
image = Image.open("file path")
instruction = "description"  # Instruction for grounding
width, height = image.width, image.height

resized_height, resized_width = smart_resize(
    image.height,
    image.width,
    factor=processor.image_processor.patch_size * processor.image_processor.merge_size,
    min_pixels=processor.image_processor.min_pixels,
    max_pixels=processor.image_processor.max_pixels,
)
resized_image = image.resize((resized_width, resized_height))
scale_x, scale_y = width / resized_width, height / resized_height

# Prepare system and user messages
system_message = {
   "role": "system",
   "content": SYSTEM_PROMPT.format(height=resized_height,width=resized_width)
}

user_message = {
    "role": "user",
    "content": [
        {"type": "image", "image": resized_image},
        {"type": "text", "text": instruction}
    ]
}

# Tokenize and prepare inputs
image_inputs, video_inputs = process_vision_info([system_message, user_message])
text = processor.apply_chat_template([system_message, user_message], tokenize=False, add_generation_prompt=True)
inputs = processor(text=[text], images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt")
inputs = inputs.to(model.device)

# Generate prediction
output_ids = model.generate(**inputs, max_new_tokens=max_new_tokens, do_sample=False, temperature=1.0, use_cache=True)
generated_ids = [output_ids[len(input_ids):] for input_ids, output_ids in zip(inputs.input_ids, output_ids)]
output_text = processor.batch_decode(generated_ids, skip_special_tokens=True, clean_up_tokenization_spaces=True)[0]

# Extract and rescale coordinates
pred_x, pred_y  = extract_coordinates(output_text) 
pred_x*=scale_x
pred_y*=scale_y 
print(pred_x,pred_y)

Refer to our code for more details.

📚 Documentation

Performance

We adhere to the standard evaluation protocol and benchmark our model on three challenging datasets. Our method consistently outperforms all other open - source model families. The following are the comparative results:

⚠️ Important Note

• Model size is indicated in billions (B) of parameters.
• A dash (—) denotes results that are currently unavailable.
• A superscript asterisk (﹡) denotes our evaluated result.
• UI-TARS-1.5 7B, Qwen2.5-VL-32B-Instruct, and Qwen2.5-VL-72B-Instruct are applied as our baseline models.
• ∆ indicates the performance improvement (∆) of our model compared to its baseline.

🔧 Technical Details

No technical details (more than 50 - word specific technical explanations) are provided in the original document, so this section is skipped.

📄 License

No license information is provided in the original document, so this section is skipped.