# Phi-Ground **Repository Path**: mirrors_microsoft/Phi-Ground ## Basic Information - **Project Name**: Phi-Ground - **Description**: No description available - **Primary Language**: Unknown - **License**: MIT - **Default Branch**: main - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 0 - **Forks**: 0 - **Created**: 2025-09-09 - **Last Updated**: 2025-09-13 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README # Phi-Ground
🤖 HomePage | 📄 Paper | 📄 Arxiv | 😊 Model | 😊 Eval data
Home page for Microsoft Phi-Ground tech-report.  ## Paper Abstract With the development of multimodal reasoning models, Computer Use Agents (CUAs), akin to Jarvis from "Iron Man", are becoming a reality. GUI grounding is a core component for CUAs to execute actual actions, similar to mechanical control in robotics, and it directly leads to the success or failure of the system. It determines actions such as clicking and typing, as well as related parameters like the coordinates for clicks. Current end-to-end grounding models still achieve less than 65% accuracy on challenging benchmarks like ScreenSpot-pro and UI-Vision, indicating they are far from being ready for deployment. In this work, we conduct an empirical study on the training of grounding models, examining details from data collection to model training. Ultimately, we developed the Phi-Ground model family, which achieves state-of-the-art performance across all five grounding benchmarks for models under 10B parameters in agent settings. In the end-to-end model setting, our model still achieves SOTA results with scores of 43.2 on ScreenSpot-pro and 27.2 on UI-Vision. We believe that the various details discussed in this paper, along with our successes and failures, not only clarify the construction of grounding models but also benefit other perception tasks. ## News - 🔥 #2 Paper of the day in huggingface [daily papers](https://huggingface.co/papers/2507.23779) ## Release Plans - [x] [Phi-Ground-4B-7C](https://huggingface.co/microsoft/Phi-Ground) - [x] Evaluation code for benchmarks - [x] [GPT-4O & O4-mini's planning data for evaluation](benchmark/new_annotations) Please stay tuned! ## How to use ### Usage The current `transformers` version can be verified with: `pip list | grep transformers`. Examples of required packages: ``` flash_attn==2.5.8 numpy==1.24.4 Pillow==10.3.0 Requests==2.31.0 torch==2.3.0 torchvision==0.18.0 transformers==4.43.0 accelerate==0.30.0 ``` ### Input Formats The model require strict input format including fixed image resolution, instruction-first order and system prompt. Input preprocessing ```python from PIL import Image def process_image(img): target_width, target_height = 336 * 3, 336 *2 img_ratio = img.width / img.height target_ratio = target_width / target_height if img_ratio > target_ratio: new_width = target_width new_height = int(new_width / img_ratio) else: new_height = target_height new_width = int(new_height * img_ratio) reshape_ratio = new_width / img.width img = img.resize((new_width, new_height), Image.LANCZOS) new_img = Image.new("RGB", (target_width, target_height), (255, 255, 255)) paste_position = (0, 0) new_img.paste(img, paste_position) return new_img instruction = "