Search is not available for this dataset
image1_jpeg unknown | image1_metadata unknown | image2_jpeg unknown | image2_metadata unknown | image1_embedding0 listlengths 1.28k 1.28k | image2_embedding0 listlengths 1.28k 1.28k |
|---|---|---|---|---|---|
"/9j/4AAQSkZJRgABAQEASABIAAD/4gxYSUNDX1BST0ZJTEUAAQEAAAxITGlubwIQAABtbnRyUkdCIFhZWiAHzgACAAkABgAxAAB(...TRUNCATED) | [
8,
1,
18,
20,
98,
99,
49,
53,
56,
49,
52,
102,
100,
55,
99,
99,
50,
98,
101,
98,
46,
106,
112,
103
] | "/9j/4AAQSkZJRgABAQEASABIAAD/2wBDAAMCAgICAgMCAgIDAwMDBAYEBAQEBAgGBgUGCQgKCgkICQkKDA8MCgsOCwkJDRENDg8(...TRUNCATED) | [
8,
1,
18,
20,
98,
49,
48,
97,
57,
97,
52,
57,
102,
48,
54,
50,
57,
50,
98,
53,
46,
106,
112,
103
] | [-0.6510127782821655,1.13981032371521,1.932199239730835,2.307011604309082,-0.7147663235664368,-1.549(...TRUNCATED) | [0.4301413595676422,0.5728847980499268,2.626241445541382,3.3087284564971924,-0.9963874816894531,2.37(...TRUNCATED) |
"/9j/4AAQSkZJRgABAQEASABIAAD/2wBDAAIBAQIBAQICAgICAgICAwUDAwMDAwYEBAMFBwYHBwcGBwcICQsJCAgKCAcHCg0KCgs(...TRUNCATED) | [
8,
1,
18,
20,
49,
57,
53,
56,
54,
101,
55,
55,
51,
55,
56,
49,
102,
99,
54,
50,
46,
106,
112,
103
] | "/9j/4AAQSkZJRgABAQEA+gD6AAD/4gxYSUNDX1BST0ZJTEUAAQEAAAxITGlubwIQAABtbnRyUkdCIFhZWiAHzgACAAkABgAxAAB(...TRUNCATED) | [
8,
1,
18,
20,
100,
101,
50,
100,
54,
102,
52,
98,
53,
52,
49,
53,
49,
54,
49,
54,
46,
106,
112,
103
] | [-0.48465654253959656,-2.228081703186035,-1.5460995435714722,1.3210842609405518,0.2881599962711334,-(...TRUNCATED) | [2.985311985015869,-1.7032053470611572,1.6645989418029785,1.1416926383972168,-0.6599526405334473,1.4(...TRUNCATED) |
"/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAIBAQEBAQIBAQECAgICAgQDAgICAgUEBAMEBgUGBgYFBgYGBwkIBgcJBwYGCAsICQo(...TRUNCATED) | [
8,
1,
18,
20,
50,
56,
100,
51,
49,
50,
57,
51,
98,
100,
49,
57,
99,
52,
56,
102,
46,
106,
112,
103
] | "/9j/4AAQSkZJRgABAQEBLAEsAAD/4gJASUNDX1BST0ZJTEUAAQEAAAIwQURCRQIQAABtbnRyUkdCIFhZWiAHzwAGAAMAAAAAAAB(...TRUNCATED) | [
8,
1,
18,
20,
50,
53,
53,
56,
102,
102,
52,
48,
97,
54,
99,
50,
97,
57,
56,
57,
46,
106,
112,
103
] | [-0.7112335562705994,-1.0397582054138184,0.007368321530520916,-0.5299071669578552,-3.520720243453979(...TRUNCATED) | [2.820969343185425,-1.4701650142669678,1.0945218801498413,-0.6924114227294922,1.288290023803711,1.24(...TRUNCATED) |
"/9j/4AAQSkZJRgABAQEASABIAAD/2wBDAAICAgICAQICAgIDAgIDAwYEAwMDAwcFBQQGCAcJCAgHCAgJCg0LCQoMCggICw8LDA0(...TRUNCATED) | [
8,
1,
18,
20,
52,
97,
56,
98,
53,
49,
100,
52,
55,
54,
51,
51,
102,
57,
100,
54,
46,
106,
112,
103
] | "/9j/4AAQSkZJRgABAQEASABIAAD/2wBDAAMCAgMCAgMDAgMDAwMDBAcFBAQEBAkGBwUHCgkLCwoJCgoMDREODAwQDAoKDhQPEBE(...TRUNCATED) | [
8,
1,
18,
20,
49,
52,
50,
99,
52,
48,
57,
102,
51,
53,
57,
56,
51,
48,
57,
100,
46,
106,
112,
103
] | [-0.027153700590133667,-0.3771643042564392,1.357200264930725,0.055514439940452576,-1.083340287208557(...TRUNCATED) | [0.8986680507659912,-0.3133907914161682,2.168400526046753,1.9164812564849854,-4.957541465759277,1.31(...TRUNCATED) |
"/9j/4AAQSkZJRgABAQAAZABkAAD/2wBDAAMCAgICAgMCAgIDAwMDBAYEBAQEBAgGBgUGCQgKCgkICQkKDA8MCgsOCwkJDRENDg8(...TRUNCATED) | [
8,
1,
18,
20,
54,
55,
97,
56,
55,
100,
98,
56,
52,
55,
54,
50,
101,
53,
53,
56,
46,
106,
112,
103
] | "/9j/4AAQSkZJRgABAQAASABIAAD/2wBDAAYEBQYFBAYGBQYHBwYIChAKCgkJChQODwwQFxQYGBcUFhYaHSUfGhsjHBYWICwgIyY(...TRUNCATED) | [
8,
1,
18,
20,
97,
53,
52,
52,
56,
51,
50,
48,
49,
49,
49,
50,
100,
54,
49,
54,
46,
106,
112,
103
] | [-2.921067714691162,1.7415411472320557,2.676335334777832,-0.3053922653198242,-1.7279365062713623,0.4(...TRUNCATED) | [0.2079678773880005,-1.0387203693389893,0.9264978766441345,-2.4495155811309814,1.8629944324493408,1.(...TRUNCATED) |
"/9j/4AAQSkZJRgABAQAAAQABAAD/4gxYSUNDX1BST0ZJTEUAAQEAAAxITGlubwIQAABtbnRyUkdCIFhZWiAHzgACAAkABgAxAAB(...TRUNCATED) | [
8,
1,
18,
20,
102,
48,
51,
49,
48,
98,
57,
56,
48,
54,
101,
53,
56,
53,
52,
48,
46,
106,
112,
103
] | "/9j/4AAQSkZJRgABAQEASABIAAD/2wBDAAICAgICAQICAgIDAgIDAwYEAwMDAwcFBQQGCAcJCAgHCAgJCg0LCQoMCggICw8LDA0(...TRUNCATED) | [
8,
1,
18,
20,
50,
102,
100,
54,
52,
56,
53,
50,
49,
99,
55,
99,
56,
99,
57,
49,
46,
106,
112,
103
] | [0.9371122121810913,-0.5399671792984009,0.8881274461746216,2.3475866317749023,-1.0516059398651123,-0(...TRUNCATED) | [-0.03058105707168579,-0.05097752809524536,1.3508915901184082,-4.207669258117676,-2.2006423473358154(...TRUNCATED) |
"/9j/4AAQSkZJRgABAQEAYABgAAD/2wBDAAMCAgMCAgMDAwMEAwMEBQgFBQQEBQoHBwYIDAoMDAsKCwsNDhIQDQ4RDgsLEBYQERM(...TRUNCATED) | [
8,
1,
18,
20,
101,
56,
99,
49,
99,
102,
99,
102,
50,
57,
98,
53,
97,
49,
53,
102,
46,
106,
112,
103
] | "/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAgGBgcGBQgHBwcJCQgKDBQNDAsLDBkSEw8UHRofHh0aHBwgJC4nICIsIxwcKDcpLDA(...TRUNCATED) | [
8,
1,
18,
20,
51,
52,
56,
102,
54,
51,
57,
97,
99,
99,
54,
48,
98,
99,
55,
98,
46,
106,
112,
103
] | [-3.210007429122925,-0.8853745460510254,2.726097583770752,-0.7139812707901001,-1.9719719886779785,2.(...TRUNCATED) | [-1.3647544384002686,1.3162680864334106,-2.338233232498169,2.3431475162506104,0.2681252360343933,1.1(...TRUNCATED) |
"/9j/4AAQSkZJRgABAQEASABIAAD/2wBDAAIBAQIBAQICAgICAgICAwUDAwMDAwYEBAMFBwYHBwcGBwcICQsJCAgKCAcHCg0KCgs(...TRUNCATED) | [
8,
1,
18,
20,
53,
53,
51,
100,
57,
97,
100,
101,
100,
57,
52,
49,
49,
48,
48,
99,
46,
106,
112,
103
] | "/9j/4AAQSkZJRgABAQEAggCCAAD/4gxYSUNDX1BST0ZJTEUAAQEAAAxITGlubwIQAABtbnRyUkdCIFhZWiAHzgACAAkABgAxAAB(...TRUNCATED) | [
8,
1,
18,
20,
101,
102,
55,
51,
49,
48,
98,
48,
50,
56,
52,
98,
100,
50,
97,
99,
46,
106,
112,
103
] | [-0.9031729102134705,0.8744885921478271,0.43783098459243774,3.1643738746643066,2.8688430786132812,-1(...TRUNCATED) | [1.941161870956421,-2.686781406402588,1.9100639820098877,-0.8338409662246704,-3.2084598541259766,2.7(...TRUNCATED) |
"/9j/4AAQSkZJRgABAQEBLAEsAAD/2wBDAAMCAgMCAgMDAwMEAwMEBQgFBQQEBQoHBwYIDAoMDAsKCwsNDhIQDQ4RDgsLEBYQERM(...TRUNCATED) | [
8,
1,
18,
20,
99,
97,
57,
98,
97,
54,
49,
49,
49,
100,
101,
55,
55,
57,
102,
54,
46,
106,
112,
103
] | "/9j/4AAQSkZJRgABAQEBLAEsAAD/4gJASUNDX1BST0ZJTEUAAQEAAAIwQURCRQIQAABtbnRyUkdCIFhZWiAHzwAGAAMAAAAAAAB(...TRUNCATED) | [
8,
1,
18,
20,
100,
48,
48,
55,
52,
55,
57,
101,
51,
50,
50,
54,
51,
52,
101,
53,
46,
106,
112,
103
] | [1.3166608810424805,0.21988607943058014,-0.6625102758407593,-0.5602558255195618,0.4014946222305298,0(...TRUNCATED) | [1.4933266639709473,-0.027342915534973145,1.8112311363220215,1.6886364221572876,0.23372256755828857,(...TRUNCATED) |
"/9j/4AAQSkZJRgABAQEAtAC0AAD/2wBDAAMCAgMCAgMDAwMEAwMEBQgFBQQEBQoHBwYIDAoMDAsKCwsNDhIQDQ4RDgsLEBYQERM(...TRUNCATED) | [
8,
1,
18,
20,
98,
100,
48,
51,
48,
56,
52,
52,
102,
53,
51,
50,
53,
97,
52,
101,
46,
106,
112,
103
] | "/9j/4AAQSkZJRgABAQEASABIAAD/2wBDAAIBAQIBAQICAgICAgICAwUDAwMDAwYEBAMFBwYHBwcGBwcICQsJCAgKCAcHCg0KCgs(...TRUNCATED) | [
8,
1,
18,
20,
56,
50,
98,
99,
97,
50,
98,
98,
55,
100,
99,
49,
52,
53,
53,
51,
46,
106,
112,
103
] | [-1.7799559831619263,-1.3368339538574219,-1.4977540969848633,-1.6588003635406494,-1.7910118103027344(...TRUNCATED) | [2.0462231636047363,0.7423720359802246,2.3731977939605713,-0.5411560535430908,0.06876523792743683,2.(...TRUNCATED) |
End of preview. Expand in Data Studio
Paired Open Images with PE-Core-G14-448 Embeddings
This dataset contains pairs of images from Open Images along with their embeddings computed using Meta's Perception Encoder (PE-Core-G14-448).
Each row contains two images (as JPEG bytes), their metadata, and their corresponding 1280-dimensional embeddings.
Data Layout
| Column | Description |
|---|---|
image1_jpeg |
JPEG bytes for the first image |
image1_metadata |
Metadata for the first image |
image2_jpeg |
JPEG bytes for the second image |
image2_metadata |
Metadata for the second image |
image1_embedding0 |
PE-Core-G14-448 embedding (dim=1280) for image 1 |
image2_embedding0 |
PE-Core-G14-448 embedding (dim=1280) for image 2 |
Splits
| Split | Files |
|---|---|
| train | 94 parquet shards |
| validation | 20 parquet shards |
| test | 5 parquet shards |
Using the Embeddings
The embeddings behave like CLIP embeddings. You can use them for zero-shot classification, retrieval, or similarity search.
Generating text embeddings for comparison
git clone https://github.com/facebookresearch/perception_models.git
cd perception_models
import torch
import core.vision_encoder.pe as pe
import core.vision_encoder.transforms as transforms
model = pe.CLIP.from_config("PE-Core-G14-448", pretrained=True).cuda().eval()
tokenizer = transforms.get_text_tokenizer(model.context_length)
text_tokens = tokenizer(["dog", "cat"]).cuda()
with torch.no_grad(), torch.autocast("cuda"):
_, text_features, _ = model(None, text_tokens)
# Compare with stored embeddings via cosine similarity
image_features = ... # load from parquet
image_features = image_features / image_features.norm(dim=-1, keepdim=True)
text_features = text_features / text_features.norm(dim=-1, keepdim=True)
similarity = model.logit_scale.exp() * image_features @ text_features.T
probs = similarity.softmax(dim=-1)
License
The images originate from Open Images, which is licensed under CC BY 4.0.
- Downloads last month
- 140