Update pipeline tag to any-to-any (#1)
Browse files- Update pipeline tag to any-to-any (332db20b081415ccd316e1f4bd644f031d783acf)
Co-authored-by: Niels Rogge <[email protected]>
README.md
CHANGED
|
@@ -1,15 +1,15 @@
|
|
| 1 |
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
library_name: transformers
|
| 4 |
base_model: Qwen/Qwen2-VL-7B-Instruct
|
| 5 |
-
|
|
|
|
|
|
|
| 6 |
---
|
| 7 |
|
| 8 |
# OS-Genesis: Automating GUI Agent Trajectory Construction via Reverse Task Synthesis
|
| 9 |
|
| 10 |
<div align="center">
|
| 11 |
|
| 12 |
-
[
|
| 13 |
|
| 14 |
</div>
|
| 15 |
|
|
@@ -21,14 +21,14 @@ We introduce OS-Genesis, an interaction-driven pipeline that synthesizes high-qu
|
|
| 21 |
## Quick Start
|
| 22 |
OS-Genesis-7B-AW is a mobile action model finetuned from [Qwen2-VL-7B-Instruct](https://huggingface.co/Qwen/Qwen2-VL-7B-Instruct).
|
| 23 |
|
| 24 |
-
###
|
| 25 |
In the following table, we provide an overview of the OS-Genesis AW Family Models used for evaluating the AndroidControl Benchmark.
|
| 26 |
|
| 27 |
-
|
|
| 28 |
| :-------------: | :-------------------------------------------------------------------------------------: | :----------------------------------------------------------------------------: | :---------------------------------------------------------: |
|
| 29 |
-
| OS-Genesis-4B-AW | [InternVL2-4B](https://huggingface.co/OpenGVLab/InternVL2-4B)
|
| 30 |
-
| OS-Genesis-7B-AW | [Qwen2-VL-7B-Instruct](https://huggingface.co/Qwen/Qwen2-VL-7B-Instruct) | [OS-Genesis-aw-training-data](https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/blob/main/os_genesis_aw_training_data.jsonl) | [🤗 link](https://huggingface.co/OS-Copilot/OS-Genesis-7B-AW)
|
| 31 |
-
| OS-Genesis-8B-AW | [InternVL2-8B](https://huggingface.co/OpenGVLab/InternVL2-8B)
|
| 32 |
|
| 33 |
|
| 34 |
### Inference Example
|
|
@@ -58,7 +58,11 @@ messages = [
|
|
| 58 |
"type": "image",
|
| 59 |
"image": "./web_6f93090a-81f6-489e-bb35-1a2838b18c01.png",
|
| 60 |
},
|
| 61 |
-
{"type": "text", "text": "You are a GUI task expert, I will provide you with a high-level instruction, an action history, a screenshot with its corresponding accessibility tree
|
|
|
|
|
|
|
|
|
|
|
|
|
| 62 |
],
|
| 63 |
}
|
| 64 |
]
|
|
|
|
| 1 |
---
|
|
|
|
|
|
|
| 2 |
base_model: Qwen/Qwen2-VL-7B-Instruct
|
| 3 |
+
library_name: transformers
|
| 4 |
+
license: apache-2.0
|
| 5 |
+
pipeline_tag: any-to-any
|
| 6 |
---
|
| 7 |
|
| 8 |
# OS-Genesis: Automating GUI Agent Trajectory Construction via Reverse Task Synthesis
|
| 9 |
|
| 10 |
<div align="center">
|
| 11 |
|
| 12 |
+
[\\[🏠Homepage\\]](https://qiushisun.github.io/OS-Genesis-Home/) [\\[💻Code\\]](https://github.com/OS-Copilot/OS-Genesis) [\\[📝Paper\\]](https://arxiv.org/abs/2412.19723) [\\[🤗Models\\]](https://huggingface.co/collections/OS-Copilot/os-genesis-6768d4b6fffc431dbf624c2d)[\\[🤗Data\\]](https://huggingface.co/collections/OS-Copilot/os-genesis-6768d4b6fffc431dbf624c2d)
|
| 13 |
|
| 14 |
</div>
|
| 15 |
|
|
|
|
| 21 |
## Quick Start
|
| 22 |
OS-Genesis-7B-AW is a mobile action model finetuned from [Qwen2-VL-7B-Instruct](https://huggingface.co/Qwen/Qwen2-VL-7B-Instruct).
|
| 23 |
|
| 24 |
+
### OS-Genesis AW Family Models
|
| 25 |
In the following table, we provide an overview of the OS-Genesis AW Family Models used for evaluating the AndroidControl Benchmark.
|
| 26 |
|
| 27 |
+
| Model Name | Base Model | Training Data | HF Link |
|
| 28 |
| :-------------: | :-------------------------------------------------------------------------------------: | :----------------------------------------------------------------------------: | :---------------------------------------------------------: |
|
| 29 |
+
| OS-Genesis-4B-AW | [InternVL2-4B](https://huggingface.co/OpenGVLab/InternVL2-4B) | [OS-Genesis-aw-training-data](https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/blob/main/os_genesis_aw_training_data.jsonl) | [🤗 link](https://huggingface.co/OS-Copilot/OS-Genesis-4B-AW) |
|
| 30 |
+
| OS-Genesis-7B-AW | [Qwen2-VL-7B-Instruct](https://huggingface.co/Qwen/Qwen2-VL-7B-Instruct) | [OS-Genesis-aw-training-data](https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/blob/main/os_genesis_aw_training_data.jsonl) | [🤗 link](https://huggingface.co/OS-Copilot/OS-Genesis-7B-AW) |
|
| 31 |
+
| OS-Genesis-8B-AW | [InternVL2-8B](https://huggingface.co/OpenGVLab/InternVL2-8B) | [OS-Genesis-aw-training-data](https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/blob/main/os_genesis_aw_training_data.jsonl) | [🤗 link](https://huggingface.co/OS-Copilot/OS-Genesis-8B-AW) |
|
| 32 |
|
| 33 |
|
| 34 |
### Inference Example
|
|
|
|
| 58 |
"type": "image",
|
| 59 |
"image": "./web_6f93090a-81f6-489e-bb35-1a2838b18c01.png",
|
| 60 |
},
|
| 61 |
+
{"type": "text", "text": "You are a GUI task expert, I will provide you with a high-level instruction, an action history, a screenshot with its corresponding accessibility tree.
|
| 62 |
+
High-level instruction: {high_level_instruction}
|
| 63 |
+
Action history: {action_history}
|
| 64 |
+
Accessibility tree: {a11y_tree}
|
| 65 |
+
Please generate the low-level thought and action for the next step."},
|
| 66 |
],
|
| 67 |
}
|
| 68 |
]
|