Skip to content

Commit aff6e33

Browse files
authored
Update README.md
1 parent 4f523cb commit aff6e33

File tree

1 file changed

+24
-12
lines changed

1 file changed

+24
-12
lines changed

Mobile-Agent-v3.5/README.md

Lines changed: 24 additions & 12 deletions
Original file line numberDiff line numberDiff line change
@@ -16,8 +16,9 @@
1616
## 📍 TODO
1717
- [x] Open source GUI-Owl 1.5 model weights
1818
- [x] Open source GUI-Owl 1.5 model cookbook
19+
- [x] Deploy Mobile-Agent-v3.5 on your own device
1920
- [ ] Open source evaluation code on benchmarks
20-
- [ ] Deploy Mobile-Agent-v3.5 on your own device
21+
2122

2223
## Introduction
2324

@@ -39,10 +40,10 @@ GUI-Owl 1.5 is the next-generation native GUI agent model family built on Qwen3-
3940
- 🤝 **Multi-agent ready**: Serves both as a standalone end-to-end agent and as specialized roles (planner, executor, verifier, notetaker) within the Mobile-Agent-v3.5 framework.
4041
-**Instruct & Thinking variants**: Smaller instruct models for fast inference and edge deployment; larger thinking models for complex tasks requiring planning and reflection.
4142

42-
<!--
43+
4344
## Deploy Mobile-Agent-v3.5 on Your Mobile Device
4445

45-
❗ At present, only **Android OS** and **Harmony OS** support tool debugging. Other systems, such as **iOS**, do not support the use of Mobile-Agent for the time being.
46+
❗ At present, only **Android OS** support tool debugging. Other systems, such as **iOS**, do not support the use of Mobile-Agent for the time being.
4647

4748
### Install Dependencies
4849
```bash
@@ -68,8 +69,8 @@ pip install numpy
6869

6970
#### Android
7071
```bash
71-
cd Mobile-Agent-v3.5/mobile_v3_5
72-
python run_mobileagentv3_5.py \
72+
cd Mobile-Agent-v3.5/mobile_use
73+
python run_gui_owl_1_5_for_mobile.py \
7374
--adb_path "Your ADB path" \
7475
--api_key "Your api key of vllm service" \
7576
--base_url "Your base url of vllm service" \
@@ -78,22 +79,33 @@ python run_mobileagentv3_5.py \
7879
--add_info "Some supplementary knowledge, can also be empty"
7980
```
8081

81-
#### HarmonyOS
82+
### Note
83+
1. GUI-Owl 1.5 outputs relative coordinates (0–1000) by default.
84+
85+
## Deploy Mobile-Agent-v3.5 on Your Computer Device
86+
87+
### Install Dependencies
8288
```bash
83-
cd Mobile-Agent-v3.5/mobile_v3_5
84-
python run_mobileagentv3_5.py \
85-
--hdc_path "Your HDC path" \
89+
pip install pyautogui
90+
pip install pyperclip
91+
```
92+
93+
### Run
94+
95+
```bash
96+
cd Mobile-Agent-v3.5/computer_use
97+
python run_gui_owl_1_5_for_pc.py \
8698
--api_key "Your api key of vllm service" \
8799
--base_url "Your base url of vllm service" \
88100
--model "Your model name of vllm service" \
89101
--instruction "The instruction you want Mobile-Agent-v3.5 to complete" \
90102
--add_info "Some supplementary knowledge, can also be empty"
91103
```
92-
-->
93104

94105
### Note
95106
1. GUI-Owl 1.5 outputs relative coordinates (0–1000) by default.
96107

108+
97109
<!--
98110
## Evaluation on OSWorld
99111
1. Follow the [official OSWorld repository](https://github.com/xlang-ai/OSWorld?tab=readme-ov-file#-installation) to install OSWorld and dependencies.
@@ -136,8 +148,8 @@ sh run_ma3_5.sh
136148
| GUI-Owl-1.5-4B-Instruct | 48.2 | 69.8 | 31.7 | 32.3 | 29.4 |
137149
| GUI-Owl-1.5-8B-Instruct | 52.3 | 69.0 | 41.8 | 41.8 | 31.7 |
138150
| GUI-Owl-1.5-8B-Thinking | 52.9 | **71.6** | 38.8 | 33.3 | 35.1 |
139-
| GUI-Owl-1.5-32B-Instruct | - | 69.4 | **47.6** | **46.8** | **44.8** |
140-
| GUI-Owl-1.5-32B-Thinking | - | 68.2 | 43.8 | 42.8 | 44.1 |
151+
| GUI-Owl-1.5-32B-Instruct | 56.5 | 69.4 | **47.6** | **46.8** | **44.8** |
152+
| GUI-Owl-1.5-32B-Thinking | 56.0 | 68.2 | 43.8 | 42.8 | 44.1 |
141153

142154
### Grounding Benchmarks
143155

0 commit comments

Comments
 (0)