Improve model card: Update paper link and add paper introduction

README.md

@@ -1,28 +1,30 @@
 ---
-license: apache-2.0
+base_model:
+- Qwen/Qwen2.5-VL-3B-Instruct
 datasets:
 - OpenGVLab/ScaleCUA-Data
 language:
 - en
+library_name: transformers
+license: apache-2.0
 metrics:
 - accuracy
-base_model:
-- Qwen/Qwen2.5-VL-3B-Instruct
 pipeline_tag: image-text-to-text
-library_name: transformers
 tags:
 - agent
 ---
 
 # SCALECUA: SCALING UP COMPUTER USE AGENTS WITH CROSS-PLATFORM DATA
 
-[\[📂 GitHub\]](https://github.com/OpenGVLab/ScaleCUA) [\[📜 Paper\]](https://github.com/OpenGVLab/ScaleCUA) [\[🚀 Quick Start\]](#model-loading)
+This model is part of the work presented in the paper [ScaleCUA: Scaling Open-Source Computer Use Agents with Cross-Platform Data](https://huggingface.co/papers/2509.15221).
+
+[\[📂 GitHub\]](https://github.com/OpenGVLab/ScaleCUA) [\[📜 Paper\]](https://huggingface.co/papers/2509.15221) [\[🚀 Quick Start\]](#model-loading)
 
 
 
 ## Introduction
 
-Recent advances in Vision-Language Models have enabled the development of agents capable of automating interactions with graphical user interfaces. Some computer use agents demonstrate strong performance, while they are typically built on closed-source models or inaccessible proprietary datasets. Moreover, the existing open-source datasets still remain insufficient for developing cross-platform general-purpose computer-use agents. To bridge this gap, we scale up the computer use dataset, constructed via a novel dual-loop interactive pipeline that combines an automated agent and a human expert into data collection. It spans **6 operating systems** and **3 task domains**, offering a large-scale and diverse corpus for training computer use agents. 
+Recent advances in Vision-Language Models (VLMs) have enabled the development of agents capable of automating interactions with graphical user interfaces. Some computer use agents demonstrate strong performance, while they are typically built on closed-source models or inaccessible proprietary datasets. Moreover, the existing open-source datasets still remain insufficient for developing cross-platform general-purpose computer-use agents. To bridge this gap, we scale up the computer use dataset, constructed via a novel dual-loop interactive pipeline that combines an automated agent and a human expert into data collection. It spans **6 operating systems** and **3 task domains**, offering a large-scale and diverse corpus for training computer use agents. 
 Building on this corpus, we develop **ScaleCUA**, capable of seamless operation across heterogeneous platforms. Trained on our dataset, it delivers consistent gains on several benchmarks, improving absolute success rates by **+26.6 points** on WebArena-Lite-v2 and **+10.7 points** on ScreenSpot-Pro compared to the baseline. Moreover, our ScaleCUA family achieves state-of-the-art performance across multiple benchmarks, e.g., **94.4%** on MMBench-GUI L1-Hard, **60.6%** on OSWorld-G and **47.4%** on WebArena-Lite-v2. These results highlight the effectiveness of our data-centric methodology in scaling both GUI understanding, grounding, and cross-platform task completion. We make our data, models, and code publicly available to facilitate future research: https://github.com/OpenGVLab/ScaleCUA.
 
 ![image/png](https://cdn-uploads.huggingface.co/production/uploads/6502f241b1792803da7e8def/YdK0I790ehLAKpR1vGkX1.png)
@@ -315,7 +317,8 @@ Previous operations:
 def format_history(history):
     if len(history) > 0:
         actions_history = [f"Step {i+1}: {low_level}" for i, low_level in enumerate(history)]
-        return "\n".join(actions_history) 
+        return "
+".join(actions_history) 
     else:
         return None
 
@@ -382,7 +385,8 @@ def parse_response(response: str) -> Dict:
     if action_matches:
         for match in action_matches:
             # Split each match by newline and strip whitespace from each line
-            lines = [line.strip() for line in match.split('\n') if line.strip()]
+            lines = [line.strip() for line in match.split('
+') if line.strip()]
             actions.extend(lines)
     operation_match = re.search(r'<operation>\s*(.*?)\s*</operation>', response, re.DOTALL)
     operation = operation_match.group(1).strip() if operation_match else None
@@ -419,7 +423,7 @@ def parse_actions(self, actions):
             keys_keyword_match = re.search(r"keys\s*=\s*(.*)", args_str, re.DOTALL)
             if keys_keyword_match:
                 keys_str = keys_keyword_match.group(1).strip()
-                if (keys_str.startswith("'") and keys_str.endswith("'")) or \
+                if (keys_str.startswith("'") and keys_str.endswith("'")) or \\
                 (keys_str.startswith('"') and keys_str.endswith('"')):
                     keys_str = keys_str[1:-1]
                 elif keys_str.startswith("[") and keys_str.endswith("]"):
@@ -428,7 +432,7 @@ def parse_actions(self, actions):
                 keys = keys_str
             elif args_str:
                 keys_str = args_str.strip()
-                if (keys_str.startswith("'") and keys_str.endswith("'")) or \
+                if (keys_str.startswith("'") and keys_str.endswith("'")) or \\
                 (keys_str.startswith('"') and keys_str.endswith('"')):
                     keys_str = keys_str[1:-1]
                 keys = keys_str
@@ -456,15 +460,15 @@ def parse_actions(self, actions):
 
         else:
             if "=" in args_str:
-                for arg in re.finditer(r"(\w+)=\[([^\]]+)\]", args_str):
+                for arg in re.finditer(r"(\w+)=\\[([^\\]]+)\\]", args_str):
                     param = arg.group(1)
                     list_str = arg.group(2)
                     
                     list_items = []
-                    for item in re.finditer(r"'([^']*)'|\"([^\"]*)\"|([^,\]]+)", list_str):
+                    for item in re.finditer(r"'([^']*)'|\"([^\"]*)\"|([^,\\]]+)", list_str):
                         val = (item.group(1) or item.group(2) or item.group(3)).strip()
                         if val:
-                            list_items.append(val.strip('"\'')) 
+                            list_items.append(val.strip('\"\'')) 
                     
                     args[param] = list_items
 
@@ -483,7 +487,7 @@ def parse_actions(self, actions):
                     elif value_str.lower() in ("true", "false"):
                         value = value_str.lower() == "true"
                     else:
-                        value = value_str.strip('"\'')
+                        value = value_str.strip('\"\'')
                     
                     args[param] = value
 
@@ -493,7 +497,7 @@ def parse_actions(self, actions):
                 for arg in re.finditer(r"'([^']*)'|\"([^\"]*)\"|([^,]+)", args_str):
                     val = (arg.group(1) or arg.group(2) or arg.group(3)).strip()
                     if val:
-                        args_list.append(val.strip('"\'')) 
+                        args_list.append(val.strip('\"\'')) 
                 
                 if args_list:
                     args["args"] = args_list