File size: 3,881 Bytes
529ac17
bb8d72b
 
 
 
 
 
 
 
 
 
 
 
529ac17
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
bb8d72b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
---
license: mit
task_categories:
- image-classification
- zero-shot-image-classification
pretty_name: RENDR
size_categories:
- 10K<n<100K
tags:
- 3d
- synthetic
- rendered
- computer-graphics
dataset_info:
  features:
  - name: image
    dtype: image
  - name: label
    dtype:
      class_label:
        names:
          '0': animals
          '1': appliances
          '2': architecture
          '3': decoration
          '4': electronics
          '5': furniture
          '6': lighting
          '7': mechanical
          '8': nature
          '9': people
          '10': tools
  splits:
  - name: train
    num_bytes: 1459477777.312
    num_examples: 23836
  - name: test
    num_bytes: 236313059.742
    num_examples: 4206
  download_size: 1584949874
  dataset_size: 1695790837.0540001
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---


# RENDR Dataset

## Dataset Description

RENDR is a large-scale synthetic dataset of rendered 3D objects across 11 object categories. The dataset contains rendered images from 3D assets sourced from BlenderKit and Haven, designed for training and evaluating computer vision models on synthetic 3D data.

## Dataset Statistics

### Split Overview

| Split | Total Images | Rendered | BlenderKit Assets | Haven Assets |
|-------|--------------|----------|-------------------|--------------|
| Train | 29,291 | 23,836 | 5,397 | 58 |
| Test | 5,929 | 4,206 | 1,701 | 22 |

### Class Distribution

| Class | Train (Rendered) | Train (BlenderKit) | Train (Haven) | Test (Rendered) | Test (BlenderKit) | Test (Haven) |
|-------|------------------|--------------------|--------------:|-----------------|-------------------|-------------:|
| Animals | 2,369 | 133 | 1 | 416 | 103 | 1 |
| Appliances | 1,966 | 388 | 5 | 346 | 150 | 2 |
| Architecture | 2,224 | 523 | 7 | 392 | 171 | 3 |
| Decoration | 2,226 | 731 | 0 | 392 | 188 | 0 |
| Electronics | 1,905 | 246 | 6 | 336 | 126 | 3 |
| Furniture | 2,154 | 1,075 | 0 | 380 | 190 | 0 |
| Lighting | 1,565 | 266 | 1 | 278 | 117 | 0 |
| Mechanical | 2,150 | 386 | 18 | 380 | 151 | 8 |
| Nature | 2,782 | 799 | 0 | 492 | 217 | 0 |
| People | 2,554 | 205 | 0 | 452 | 136 | 0 |
| Tools | 1,941 | 645 | 20 | 342 | 152 | 5 |

## Dataset Structure

```
rendr/
├── train/
│   ├── animals/
│   ├── appliances/
│   ├── architecture/
│   ├── decoration/
│   ├── electronics/
│   ├── furniture/
│   ├── lighting/
│   ├── mechanical/
│   ├── nature/
│   ├── people/
│   └── tools/
└── test/
    └── [same structure as train]
```

## Normalization Statistics

For standard ImageNet-style normalization:

- **Mean**: `[0.5910, 0.5846, 0.5790]`
- **Std**: `[0.2724, 0.2733, 0.2781]`

## Usage

```python
from datasets import load_dataset

# Load the dataset
dataset = load_dataset("jneuendorf/rendr")

# Access splits
train_data = dataset['train']
test_data = dataset['test']

# Example: Load with normalization
from torchvision import transforms

transform = transforms.Compose([
    transforms.ToTensor(),
    transforms.Normalize(
        mean=[0.5910, 0.5846, 0.5790],
        std=[0.2724, 0.2733, 0.2781]
    )
])
```

## Data Sources

- **Rendered Images**: Custom rendered synthetic images
- **BlenderKit**: 3D assets from BlenderKit library
- **Haven**: 3D assets from Poly Haven

## Classes

The dataset includes 11 object categories:
1. Animals
2. Appliances
3. Architecture
4. Decoration
5. Electronics
6. Furniture
7. Lighting
8. Mechanical
9. Nature
10. People
11. Tools

## Citation

If you use this dataset, please cite:

```bibtex
@dataset{rendr,
  title={RENDR: A Large-Scale Synthetic 3D Object Dataset},
  author={Jim Neuendorf},
  year={2025}
}
```

## License

MIT License - Copyright (c) 2025 Jim Neuendorf