Statements (28)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
image-text dataset
|
| gptkbp:availableOn |
https://laion.ai/blog/laion-5b/
|
| gptkbp:citation |
gptkb:Schuhmann_et_al.,_2022
|
| gptkbp:contains |
over 5 billion image-text pairs
|
| gptkbp:createdBy |
gptkb:LAION
|
| gptkbp:dataSource |
gptkb:Common_Crawl
|
| gptkbp:filtering |
NSFW filtering
image quality filtering language filtering |
| gptkbp:format |
Parquet
|
| https://www.w3.org/2000/01/rdf-schema#label |
LAION-5B
|
| gptkbp:includes |
CLIP embeddings
image URLs text captions |
| gptkbp:language |
multilingual
|
| gptkbp:license |
gptkb:Creative_Common
|
| gptkbp:notableFor |
training OpenCLIP
training Stable Diffusion |
| gptkbp:openSource |
true
|
| gptkbp:relatedTo |
gptkb:LAION-2B-en
gptkb:LAION-400M gptkb:LAION-Aesthetics |
| gptkbp:releaseYear |
2022
|
| gptkbp:size |
240TB uncompressed
|
| gptkbp:usedFor |
gptkb:machine_learning
training large vision-language models |
| gptkbp:bfsParent |
gptkb:Stable_Diffusion
|
| gptkbp:bfsLayer |
5
|