GUI-Libra
Collection
Training GUI agents with augmented reasoning data and a tailored post-training recipe • 5 items • Updated
data_source stringclasses 2 values | images listlengths 1 1 | extra_info dict | answer dict | system_prompt stringclasses 2 values | context stringlengths 171 1.62k |
|---|---|---|---|---|---|
android_control | [
{
"bytes": [
137,
80,
78,
71,
13,
10,
26,
10,
0,
0,
0,
13,
73,
72,
68,
82,
0,
0,
4,
56,
0,
0,
9,
96,
8,
6,
0,
0,
0,
109,
26,
... | {
"img_name": "1024/screenshot_7.png",
"step_id": 8
} | {
"gt_action": "Click",
"gt_bbox": [
25,
196,
972,
519
],
"gt_input_text": "None",
"gt_point_2d": [
0.4981481481481482,
0.3575
],
"gt_target": "None",
"image_height": 2400,
"image_width": 1080
} | You are a GUI agent. You are given a task and a screenshot of the screen. You need to perform a series of actions to complete the task. You need to choose actions from the the following list:
action_type: Answer, action_target: None, value: Answer text, point_2d: None
## Explanation: Return the final answer to the user's question
action_type: Click, action_target: Element description, value: None, point_2d: [x, y]
## Explanation: Tap or click a specific UI element and provide its coordinates
action_type: Write, action_target: Element description or None, value: Text to enter, point_2d: [x, y] or None
## Explanation: Enter text into a specific input field or at the current focus if coordinate is None
action_type: LongPress, action_target: Element description, value: None, point_2d: [x, y]
## Explanation: Press and hold on a specific UI element (mobile only) and provide its coordinates
action_type: Scroll, action_target: None, value: "up" | "down" | "left" | "right", point_2d: None
## Explanation: Scroll a view or container in the specified direction
action_type: Swipe, action_target: Optional position or None, value: "up" | "down" | "left" | "right", point_2d: [x, y] or None
## Explanation: Perform a swipe gesture on the screen in the given direction (mobile)
action_type: Wait, action_target: None, value: Number of seconds, point_2d: None
## Explanation: Pause execution to allow the UI to load or update
action_type: NavigateHome, action_target: None, value: None, point_2d: None
## Explanation: Navigate to the device's home screen
action_type: NavigateBack, action_target: None, value: None, point_2d: None
## Explanation: Press the system "Back" button
action_type: OpenApp, action_target: None, value: App name, point_2d: None
## Explanation: Launch an app by its name (mobile only)
action_type: Terminate, action_target: None, value: End-task message, point_2d: None
## Explanation: Signal the end of the current task with a final message
| <image>
Please generate the next move according to the UI screenshot, instruction and previous actions.
Instruction: In the Arts & Culture app, I want to create an art gallery with the title Self Art.
Interaction History:
Step 1 Action: Go back to the previous screen
Step 2 Action: Go back to the previous screen
Step 3 Action: Swipe up to Favorites option
Step 4 Action: Swipe up to Favorites option
Step 5 Action: Click on the Favorites
Step 6 Action: Click on the Galleries section
Step 7 Action: Click on the Create Gallery button |
android_control | [{"bytes":"iVBORw0KGgoAAAANSUhEUgAABDgAAAlgCAYAAABtGofVAAABLGlDQ1BJQ0MgUHJvZmlsZQAAeJx9kb1Lw0AYxn+WL(...TRUNCATED) | {
"img_name": "7878/screenshot_3.png",
"step_id": 4
} | {"gt_action":"Click","gt_bbox":[38,254,381,409],"gt_input_text":"None","gt_point_2d":[0.209259259259(...TRUNCATED) | "You are a GUI agent. You are given a task and a screenshot of the screen. You need to perform a ser(...TRUNCATED) | "<image>\nPlease generate the next move according to the UI screenshot, instruction and previous act(...TRUNCATED) |
android_control | [{"bytes":"iVBORw0KGgoAAAANSUhEUgAABDgAAAlgCAYAAABtGofVAAABLGlDQ1BJQ0MgUHJvZmlsZQAAeJx9kb1Lw0AYxn+WL(...TRUNCATED) | {
"img_name": "4662/screenshot_1.png",
"step_id": 2
} | {"gt_action":"Click","gt_bbox":[875,60,973,113],"gt_input_text":"None","gt_point_2d":[0.924074074074(...TRUNCATED) | "You are a GUI agent. You are given a task and a screenshot of the screen. You need to perform a ser(...TRUNCATED) | "<image>\nPlease generate the next move according to the UI screenshot, instruction and previous act(...TRUNCATED) |
android_control | [{"bytes":"iVBORw0KGgoAAAANSUhEUgAABDgAAAlgCAYAAABtGofVAAABLGlDQ1BJQ0MgUHJvZmlsZQAAeJx9kb1Lw0AYxn+WL(...TRUNCATED) | {
"img_name": "7961/screenshot_1.png",
"step_id": 2
} | {"gt_action":"Click","gt_bbox":[38,377,425,422],"gt_input_text":"None","gt_point_2d":[0.231481481481(...TRUNCATED) | "You are a GUI agent. You are given a task and a screenshot of the screen. You need to perform a ser(...TRUNCATED) | "<image>\nPlease generate the next move according to the UI screenshot, instruction and previous act(...TRUNCATED) |
android_control | [{"bytes":"iVBORw0KGgoAAAANSUhEUgAABDgAAAlgCAYAAABtGofVAAABLGlDQ1BJQ0MgUHJvZmlsZQAAeJx9kb1Lw0AYxn+WL(...TRUNCATED) | {
"img_name": "9231/screenshot_0.png",
"step_id": 1
} | {"gt_action":"Click","gt_bbox":[0,524,1000,673],"gt_input_text":"None","gt_point_2d":[0.5,0.59875],"(...TRUNCATED) | "You are a GUI agent. You are given a task and a screenshot of the screen. You need to perform a ser(...TRUNCATED) | "<image>\nPlease generate the next move according to the UI screenshot, instruction and previous act(...TRUNCATED) |
android_control | [{"bytes":"iVBORw0KGgoAAAANSUhEUgAABDgAAAlgCAYAAABtGofVAAABLGlDQ1BJQ0MgUHJvZmlsZQAAeJx9kb1Lw0AYxn+WL(...TRUNCATED) | {
"img_name": "6405/screenshot_18.png",
"step_id": 18
} | {"gt_action":"Click","gt_bbox":[56,852,192,913],"gt_input_text":"None","gt_point_2d":[0.125,0.899583(...TRUNCATED) | "You are a GUI agent. You are given a task and a screenshot of the screen. You need to perform a ser(...TRUNCATED) | "<image>\nPlease generate the next move according to the UI screenshot, instruction and previous act(...TRUNCATED) |
android_control | [{"bytes":"iVBORw0KGgoAAAANSUhEUgAABDgAAAlgCAYAAABtGofVAAABLGlDQ1BJQ0MgUHJvZmlsZQAAeJx9kb1Lw0AYxn+WL(...TRUNCATED) | {
"img_name": "6903/screenshot_1.png",
"step_id": 2
} | {"gt_action":"Click","gt_bbox":[58,865,107,887],"gt_input_text":"None","gt_point_2d":[0.069444444444(...TRUNCATED) | "You are a GUI agent. You are given a task and a screenshot of the screen. You need to perform a ser(...TRUNCATED) | "<image>\nPlease generate the next move according to the UI screenshot, instruction and previous act(...TRUNCATED) |
android_control | [{"bytes":"iVBORw0KGgoAAAANSUhEUgAABDgAAAlgCAYAAABtGofVAAABLGlDQ1BJQ0MgUHJvZmlsZQAAeJx9kb1Lw0AYxn+WL(...TRUNCATED) | {
"img_name": "15409/screenshot_2.png",
"step_id": 3
} | {"gt_action":"Click","gt_bbox":[146,135,346,161],"gt_input_text":"None","gt_point_2d":[0.24629629629(...TRUNCATED) | "You are a GUI agent. You are given a task and a screenshot of the screen. You need to perform a ser(...TRUNCATED) | "<image>\nPlease generate the next move according to the UI screenshot, instruction and previous act(...TRUNCATED) |
android_control | [{"bytes":"iVBORw0KGgoAAAANSUhEUgAABDgAAAlgCAYAAABtGofVAAABLGlDQ1BJQ0MgUHJvZmlsZQAAeJx9kb1Lw0AYxn+WL(...TRUNCATED) | {
"img_name": "6404/screenshot_3.png",
"step_id": 4
} | {"gt_action":"Click","gt_bbox":[747,846,961,898],"gt_input_text":"None","gt_point_2d":[0.85370370370(...TRUNCATED) | "You are a GUI agent. You are given a task and a screenshot of the screen. You need to perform a ser(...TRUNCATED) | "<image>\nPlease generate the next move according to the UI screenshot, instruction and previous act(...TRUNCATED) |
android_control | [{"bytes":"iVBORw0KGgoAAAANSUhEUgAABDgAAAlgCAYAAABtGofVAAABLGlDQ1BJQ0MgUHJvZmlsZQAAeJx9kb1Lw0AYxn+WL(...TRUNCATED) | {
"img_name": "15865/screenshot_3.png",
"step_id": 4
} | {"gt_action":"Click","gt_bbox":[856,438,973,491],"gt_input_text":"None","gt_point_2d":[0.91481481481(...TRUNCATED) | "You are a GUI agent. You are given a task and a screenshot of the screen. You need to perform a ser(...TRUNCATED) | "<image>\nPlease generate the next move according to the UI screenshot, instruction and previous act(...TRUNCATED) |
End of preview. Expand
in Data Studio
README.md exists but content is empty.
- Downloads last month
- 14
Size of downloaded dataset files:
1.15 GB
Size of the auto-converted Parquet files:
1.15 GB
Number of rows:
738