Upload BlenderStudio dataset

.gitattributes

@@ -57,3 +57,30 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level1/camera1/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level1/camera2/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level1/camera3/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level1/camera4/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level1/camera5/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level1/camera6/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level1/camera7/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level1/camera8/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level1/camera9/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level2/attribute1/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level2/attribute2/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level2/attribute3/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level2/attribute4/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level2/attribute5/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level2/attribute6/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level2/attribute7/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level2/attribute8/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level2/attribute9/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level3/attribute1/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level3/attribute2/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level3/attribute3/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level3/attribute4/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level3/attribute5/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level3/attribute6/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level3/attribute7/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level3/attribute8/blender_file.blend filter=lfs diff=lfs merge=lfs -text
+data/blenderbench/level3/attribute9/blender_file.blend filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,251 @@
+---
+language:
+- en
+license: cc-by-4.0
+size_categories:
+- n<1K
+task_categories:
+- image-to-text
+- text-generation
+- visual-question-answering
+pretty_name: BlenderBench
+tags:
+- 3d
+- blender
+- code-generation
+- vision
+- multimodal
+- benchmark
+dataset_info:
+  features:
+  - name: instance_id
+    dtype: string
+  - name: level
+    dtype: string
+  - name: instance_name
+    dtype: string
+  - name: task_description
+    dtype: string
+  - name: start_code
+    dtype: string
+  - name: goal_code
+    dtype: string
+  - name: start_render
+    dtype: image
+  - name: goal_render
+    dtype: image
+  - name: blend_file_path
+    dtype: string
+  - name: blend_file_size_mb
+    dtype: float64
+  splits:
+  - name: train
+    num_bytes: 18000000000
+    num_examples: 27
+  download_size: 18000000000
+  dataset_size: 18000000000
+---
+
+# BlenderBench Dataset
+
+<p align="center">
+    <a href="https://fugtemypt123.github.io/VIGA-website/"><img src="https://img.shields.io/badge/Page-Project-blue" alt="Project Page"></a>
+    <a href="https://arxiv.org/abs/2601.11109"><img src="https://img.shields.io/badge/Paper-arXiv-b31b1b" alt="arXiv Paper"></a>
+    <a href="https://github.com/Fugtemypt123/VIGA-release"><img src="https://img.shields.io/badge/Code-Github-black" alt="Github Code"></a>
+</p>
+
+## Dataset Description
+
+**BlenderBench** is a comprehensive benchmark dataset for evaluating models on 3D scene editing tasks in Blender. The dataset challenges agents to understand visual differences between initial and target scenes, then generate appropriate Blender Python code to transform the initial scene to match the target.
+
+### Key Features
+
+- **27 instances** across 3 difficulty levels
+- **Multi-modal**: Combines visual understanding with code generation
+- **Realistic 3D scenes**: Using Blender Studio assets
+- **Progressive difficulty**: From simple camera adjustments to complex scene manipulations
+- **Rich annotations**: Each instance includes task descriptions, start/goal code, and rendered images
+
+### Task Categories
+
+The dataset covers three main difficulty levels:
+
+1. **Level 1** (9 instances): Camera adjustments
+2. **Level 2** (9 instances): Multi-step editing
+3. **Level 3** (9 instances): Compositional editing (level 1 + level 2)
+
+## Dataset Structure
+
+Each instance in the dataset contains:
+
+```
+instance/
+├── blender_file.blend       # Blender scene file
+├── start.py                  # Initial scene configuration code
+├── goal.py                   # Target scene configuration code
+├── task.txt                  # Natural language task description
+└── renders/
+    ├── start/
+    │   └── render1.png      # Rendered image of initial scene (512x512)
+    └── goal/
+        └── render1.png      # Rendered image of target scene (512x512)
+```
+
+### Data Fields
+
+When loaded using `datasets.load_dataset()`, each example contains:
+
+- `instance_id` (string): Unique identifier (e.g., "level1/camera1")
+- `level` (string): Difficulty level ("level1", "level2", or "level3")
+- `instance_name` (string): Instance name within the level
+- `task_description` (string): Natural language description of the task
+- `start_code` (string): Python code for the initial scene setup
+- `goal_code` (string): Python code for the target scene configuration
+- `start_render` (image): Rendered image of the initial scene (512x512 PNG)
+- `goal_render` (image): Rendered image of the target scene (512x512 PNG)
+- `blend_file_path` (string): Relative path to the .blend file
+- `blend_file_size_mb` (float): Size of the .blend file in MB
+
+### Data Splits
+
+The dataset provides a single training split containing all instances. You can filter by difficulty level using the dataset configurations:
+
+- `all`: All instances (default)
+- `level1`: Only level 1 instances
+- `level2`: Only level 2 instances
+- `level3`: Only level 3 instances
+
+
+## Usage
+
+### Installation
+
+```bash
+pip install datasets huggingface_hub
+```
+
+### Loading the Dataset
+
+```python
+from datasets import load_dataset
+
+# Load all instances
+dataset = load_dataset("DietCoke4671/BlenderBench")
+
+# Load only level 1 instances
+dataset = load_dataset("DietCoke4671/BlenderBench", "level1")
+
+# Access an example
+example = dataset["train"][0]
+print(f"Task: {example['task_description']}")
+print(f"Level: {example['level']}")
+
+# Display images
+example['start_render'].show()  # Initial scene
+example['goal_render'].show()   # Target scene
+
+# Access code
+print(f"Start code:\n{example['start_code']}")
+print(f"Goal code:\n{example['goal_code']}")
+```
+
+### Using with Blender
+
+To actually render and evaluate generated code, you'll need Blender installed:
+
+```python
+import subprocess
+from pathlib import Path
+
+def render_with_blender(blend_file, code, output_dir):
+    """
+    Execute Blender code and render the result.
+
+    Args:
+        blend_file: Path to .blend file
+        code: Python code to execute in Blender
+        output_dir: Directory to save rendered output
+    """
+    # Save code to temporary file
+    code_file = Path(output_dir) / "temp_code.py"
+    with open(code_file, 'w') as f:
+        f.write(code)
+
+    # Run Blender
+    cmd = [
+        "blender",
+        "--background",
+        str(blend_file),
+        "--python", str(code_file),
+        "--render-output", str(output_dir / "render.png"),
+        "--render-frame", "1"
+    ]
+
+    subprocess.run(cmd)
+
+# Example usage
+example = dataset["train"][0]
+render_with_blender(
+    blend_file=f"path/to/{example['blend_file_path']}",
+    code=example['start_code'],
+    output_dir="output/"
+)
+```
+
+### Example: Building an AI Agent
+
+```python
+from datasets import load_dataset
+import openai
+
+# Load dataset
+dataset = load_dataset("DietCoke4671/BlenderBench", "level1")
+
+def solve_blender_task(example):
+    """
+    Use an AI agent to solve a BlenderBench task.
+    """
+    # Prepare prompt with task description and images
+    prompt = f"""
+You are an expert Blender Python programmer. Your task is to:
+
+Task: {example['task_description']}
+
+Given the initial scene code and rendered image, generate Blender Python code
+to transform the scene to match the target image.
+
+Initial code:
+{example['start_code']}
+
+Generate the modified code that will produce the target scene.
+"""
+
+    # Call your AI model (e.g., GPT-4 with vision)
+    response = openai.ChatCompletion.create(
+        model="gpt-4-vision-preview",
+        messages=[
+            {
+                "role": "user",
+                "content": [
+                    {"type": "text", "text": prompt},
+                    {
+                        "type": "image_url",
+                        "image_url": {"url": encode_image(example['start_render'])}
+                    },
+                    {
+                        "type": "image_url",
+                        "image_url": {"url": encode_image(example['goal_render'])}
+                    }
+                ]
+            }
+        ]
+    )
+
+    return response.choices[0].message.content
+
+# Solve all level 1 tasks
+for example in dataset["train"]:
+    solution = solve_blender_task(example)
+    print(f"Solution for {example['instance_id']}:")
+    print(solution)
+```

data/blenderbench/level1/camera1/blender_file.blend

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c51ba3000030b7cf17813095327a56860055ce32fbd651d54db6c872e9f3ba2
+size 11941723

data/blenderbench/level1/camera1/goal.py

@@ -0,0 +1,4 @@
+import bpy
+
+bpy.data.objects["Camera1"].location = (-1.0292, 1.6960, 1.1705)
+bpy.data.objects["Camera1"].rotation_euler = (1.5701, 0.0000, 3.6424)

data/blenderbench/level1/camera1/start.py

@@ -0,0 +1,4 @@
+import bpy
+
+bpy.data.objects["Camera1"].location = (1.0765, -1.2848, 1.4343)
+bpy.data.objects["Camera1"].rotation_euler = (1.5701, 0.0000, 0.6753)

data/blenderbench/level1/camera1/task.txt

@@ -0,0 +1 @@
+Adjust the camera position so that the viewing angle is consistent with the target image

data/blenderbench/level1/camera2/blender_file.blend

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e53b677deab9a576b1d311f716c311def91591db41a33518a0433bc4cef8d3b
+size 21287404

data/blenderbench/level1/camera2/goal.py

@@ -0,0 +1,4 @@
+import bpy
+
+bpy.data.objects["Camera1"].location = (3.9689, -4.6158, 2.1883)
+bpy.data.objects["Camera1"].rotation_euler = (1.3188, 0.0000, 0.7800)

data/blenderbench/level1/camera2/start.py

@@ -0,0 +1,4 @@
+import bpy
+
+bpy.data.objects["Camera1"].location = (-3.4813, 4.5001, 0.5115)
+bpy.data.objects["Camera1"].rotation_euler = (1.6678, 0.0000, 3.8693)

data/blenderbench/level1/camera2/task.txt

@@ -0,0 +1 @@
+Adjust the camera position so that the viewing angle is consistent with the target image

data/blenderbench/level1/camera3/blender_file.blend

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2d8bb6bc5b8614ab26d6319249c4cf915e5e8107f7859ec768dfb0e02b875403
+size 19714277

data/blenderbench/level1/camera3/goal.py

@@ -0,0 +1,5 @@
+import bpy
+
+bpy.data.objects["Camera1"].location = (-1.6577, -1.7238, 1.9317)
+bpy.data.objects["Camera1"].rotation_euler = (1.4033, -0.0000, 5.6653)
+bpy.data.objects["Camera1"].data.lens = 21

data/blenderbench/level1/camera3/start.py

@@ -0,0 +1,5 @@
+import bpy
+
+bpy.data.objects["Camera1"].location = (6.7303, -4.8700, 3.7385)
+bpy.data.objects["Camera1"].rotation_euler = (1.2811, -0.0000, 0.9355)
+bpy.data.objects["Camera1"].data.lens = 50

data/blenderbench/level1/camera3/task.txt

@@ -0,0 +1 @@
+Adjust the camera position so that the viewing angle is consistent with the target image

data/blenderbench/level1/camera4/blender_file.blend

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e99ccd5e43e80d8dbdd4d840ea4290f1dbc6b4e06d21be07eeed7d39fecafd0a
+size 1092706714

data/blenderbench/level1/camera4/goal.py

@@ -0,0 +1,4 @@
+import bpy
+
+bpy.data.objects["Camera1"].location = (-1.1222, -0.3259, 4.4354)
+bpy.data.objects["Camera1"].rotation_euler = (-0.0873, 0.5934, 0.2094)

data/blenderbench/level1/camera4/start.py

@@ -0,0 +1,4 @@
+import bpy
+
+bpy.data.objects["Camera1"].location = (1.1400, 0.8500, 0.5300)
+bpy.data.objects["Camera1"].rotation_euler = (2.4435, 1.1170, 2.6354)

data/blenderbench/level1/camera4/task.txt

@@ -0,0 +1 @@
+Adjust the camera position so that the viewing angle is consistent with the target image

data/blenderbench/level1/camera5/blender_file.blend

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:06f30a3ccd0d81be515f1892605fd575f6933052e15b9b297b7d96a48aa664d2
+size 1092786654

data/blenderbench/level1/camera5/goal.py

@@ -0,0 +1,4 @@
+import bpy
+
+bpy.data.objects["Camera1"].location = (-3.1800, -0.0500, 2.5700)
+bpy.data.objects["Camera1"].rotation_euler = (0.3491, -0.8029, -0.2269)

data/blenderbench/level1/camera5/start.py

@@ -0,0 +1,4 @@
+import bpy
+
+bpy.data.objects["Camera1"].location = (-1.2250, 0.7100, -2.8300)
+bpy.data.objects["Camera1"].rotation_euler = (-0.4363, 3.3161, 0.0349)

data/blenderbench/level1/camera5/task.txt

@@ -0,0 +1 @@
+Adjust the camera position so that the viewing angle is consistent with the target image

data/blenderbench/level1/camera6/blender_file.blend

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e0fb03e37b9072c9ca65bab2b6594deb4254d21b3e13c08b8eeaf1fce76e45f
+size 5718592

data/blenderbench/level1/camera6/goal.py

@@ -0,0 +1,4 @@
+import bpy
+
+bpy.data.objects["Camera1"].location = (-1.0800, 1.0600, 0.7700)
+bpy.data.objects["Camera1"].rotation_euler = (1.5708, 0.0000, 3.9968)

data/blenderbench/level1/camera6/start.py

@@ -0,0 +1,4 @@
+import bpy
+
+bpy.data.objects["Camera1"].location = (2.4689, -2.6258, 1.3083)
+bpy.data.objects["Camera1"].rotation_euler = (1.4060, 0.0000, 0.7626)

data/blenderbench/level1/camera6/task.txt

@@ -0,0 +1 @@
+Adjust the camera position so that the viewing angle is consistent with the target image

data/blenderbench/level1/camera7/blender_file.blend

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6106fe112e0bfdf4e7aab26ff4e901ad8f942a9e6493839a9d795806cf97ab60
+size 198756884

data/blenderbench/level1/camera7/goal.py

@@ -0,0 +1,4 @@
+import bpy
+
+bpy.data.objects["Camera1"].location = (2.4674, 0.2196, -1.2138)
+bpy.data.objects["Camera1"].rotation_euler = (-0.4014, 2.0071, 0.0000)

data/blenderbench/level1/camera7/start.py

@@ -0,0 +1,4 @@
+import bpy
+
+bpy.data.objects["Camera1"].location = (1.3353, 1.4740, 1.5357)
+bpy.data.objects["Camera1"].rotation_euler = (-0.7679, 0.2094, 0.0000)

data/blenderbench/level1/camera7/task.txt

@@ -0,0 +1 @@
+Adjust the camera position so that the viewing angle is consistent with the target image

data/blenderbench/level1/camera8/blender_file.blend

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:64124da51512f689ff1c2d0a6475ccf377bb94a87a3bcc9fd14e0d38de7a4e00
+size 2303572

data/blenderbench/level1/camera8/goal.py

@@ -0,0 +1,4 @@
+import bpy
+
+bpy.data.objects["Camera1"].location = (-4.1656, 0.0020, 3.7237)
+bpy.data.objects["Camera1"].rotation_euler = (1.5708, 0.0000, -1.5708)

data/blenderbench/level1/camera8/start.py

@@ -0,0 +1,4 @@
+import bpy
+
+bpy.data.objects["Camera1"].location = (7.3589, -6.9258, 4.9583)
+bpy.data.objects["Camera1"].rotation_euler = (1.2664, 0.0000, 0.7800)

data/blenderbench/level1/camera8/task.txt

@@ -0,0 +1 @@
+Adjust the camera position so that the viewing angle is consistent with the target image