l

CLIP ViT L 14 DataComp.XL S13B B90K

Apr 2023
lCLIP ViT L 14 DataComp.XL S13B B90K
Compare in
on
self-host onlyprice
30.8MIEB lite
Apr 2023released
768Dimensions
0K tokensMax input

Ranking?

Full ranking →
1EVA02 CLIP B 16862Siglip So400m Patch14 38484.63Siglip Large Patch16 38483.74CLIP ViT BigG 14 Laion2B 39B B160k83.65CLIP ViT G 14 Laion2B S34B B88K82.8
···
7Siglip Large Patch16 25682.58CLIP ViT L 14 DataComp.XL S13B B90K829CLIP ViT L 14 Laion2B S32B B82K81.2

Strong at

C
Embeddings & Retrieval#65 of 89

Alternatives?

ModelScoreCostPriceWhat changes
QBetterEVA02 CLIP B 16unstatedOpen+5%——Memory (4-bit): 0.3 GB → 3 GBBetterSiglip So400m Patch14 384Open+3%——Memory (4-bit): 0.3 GB → 0.5 GBOpenSiglip Large Patch16 384+2%——Memory (4-bit): 0.3 GB → 0.4 GBlOpenCLIP ViT BigG 14 Laion2B 39B B160k+2%——Memory (4-bit): 0.3 GB → 1.5 GBlOpenCLIP ViT G 14 Laion2B S34B B88Ksame——Memory (4-bit): 0.3 GB → 0.8 GB

Run it yourself

estimate
Fits on✓ 8 GB (phone, laptop) · full quality✓ 16 GB · full quality✓ 24 GB GPU · full quality✓ 48 GB GPU · full quality✓ 80 GB GPU · full quality
LicenceMITfree for commercial use

Benchmarks

in Embeddings & Retrieval
MIEB English · Mean over tasks · multilingual (13 languages)59.4#8
MIEB English · Mean over task types · multilingual (13 languages)59.4#8
MIEB English · Document Understanding · English38.6#57
MIEB English · Zero-shot Classification · English68.4#4
MIEB English · Any-to-any Retrieval · English38.1#11
MIEB English · Visual semantic similarity (English) · multilingual (13 languages)69.9#10
MIEB English · Image Classification · English82#8
MIEB English · Compositionality · English39.1#26
MIEB English · Vision Centric QA · English52.3#13
MIEB English · Image Clustering · English86.4#2
MIEB Img · Mean over tasks · multilingual (13 languages)63.6#6
MIEB Img · Mean over task types · multilingual (13 languages)63.6#6
MIEB Img · Image Classification · English82#8
MIEB Img · Any-to-any Retrieval · English39.8#8
MIEB Img · Visual semantic similarity (English) · English69#10
MIEB Img · Visual semantic similarity (multilingual) · multilingual (13 languages)40.7#21
MIEB Img · Image Clustering · English86.4#2
MIEB lite · Mean over tasks · multilingual (41 languages)50.4#20
MIEB lite · Mean over task types · multilingual (41 languages)50.4#20
MIEB lite · Vision Centric QA · English54.9#18
MIEB lite · Any-to-any Retrieval · English31#17
MIEB lite · Document Understanding · English30.8#65
MIEB lite · Zero-shot Classification · English69.4#8
MIEB lite · Compositionality · English31.7#32
MIEB lite · Image Classification · English75.3#9
MIEB lite · Any-to-any Multilingual Retrieval · multilingual (40 languages)22.6#28
MIEB lite · Image Clustering · English80.4#7
MIEB lite · Visual semantic similarity (English) · English72.5#15
MIEB lite · Visual semantic similarity (multilingual) · multilingual (13 languages)35.8#27
MIEB multilingual · Mean over tasks · multilingual (41 languages)53.4#13
MIEB multilingual · Mean over task types · multilingual (41 languages)53.4#13
MIEB multilingual · Zero-shot Classification · English68.4#4
MIEB multilingual · Image Classification · English82#8
MIEB multilingual · Document Understanding · English38.6#57
MIEB multilingual · Any-to-any Retrieval · English38.1#11
MIEB multilingual · Compositionality · English39.1#26
MIEB multilingual · Vision Centric QA · English52.3#13
MIEB multilingual · Visual semantic similarity (English) · multilingual (13 languages)69.9#10
MIEB multilingual · Visual semantic similarity (multilingual) · multilingual (13 languages)35.8#27
MIEB multilingual · Any-to-any Multilingual Retrieval · multilingual (40 languages)23.8#20
MIEB multilingual · Image Clustering · English86.4#2

FAQ

How much does CLIP ViT L 14 DataComp.XL S13B B90K cost?›

There is no per-unit API price: self-host only.

Is CLIP ViT L 14 DataComp.XL S13B B90K free to use?›

There is no free API. The weights are open, so you can run it yourself.

How good is CLIP ViT L 14 DataComp.XL S13B B90K, and what is it best at?›

#65 of 89 in Multimodal.

Is CLIP ViT L 14 DataComp.XL S13B B90K better than Siglip Large Patch16 256?›

On MIEB Img · Image Classification · English, CLIP ViT L 14 DataComp.XL S13B B90K scores 82 vs 82.5 — about the same.

When was CLIP ViT L 14 DataComp.XL S13B B90K released? Is there a newer version?›

Released 2023-04-26. It is the current version in its line.

Is CLIP ViT L 14 DataComp.XL S13B B90K open source?›

Yes, the weights are open (MIT licence).

Can I run CLIP ViT L 14 DataComp.XL S13B B90K locally?›

The weights are open and can be self-hosted; hardware needs are not confirmed yet.

What is CLIP ViT L 14 DataComp.XL S13B B90K?›

CLIP ViT L 14 DataComp.XL S13B B90K is an open-weights embeddings & retrieval model from laion, released Apr 2023.

What are the best alternatives to CLIP ViT L 14 DataComp.XL S13B B90K?›

EVA02 CLIP B 16 (better).