Dataset Viewer
Auto-converted to Parquet Duplicate
objectID
int64
34
955k
embedding
listlengths
1.28k
1.28k
model
stringclasses
1 value
dim
int32
1.28k
1.28k
34
[ 0.023432942107319832, -0.02650412917137146, -0.0034973707515746355, 0.034308988600969315, -0.012943293899297714, -0.01154786255210638, 0.0213041752576828, -0.007750748191028833, -0.037800777703523636, -0.02226492576301098, 0.02590986341238022, -0.057879507541656494, 0.013686470687389374, 0...
laion/CLIP-ViT-bigG-14-laion2b_s39b_b160k
1,280
35
[ 0.02198607102036476, -0.0016596257919445634, -0.013230178505182266, 0.017936745658516884, -0.004672721028327942, -0.01079269777983427, 0.03551117330789566, -0.027330871671438217, 0.019556937739253044, 0.013145489618182182, 0.018768196925520897, -0.05242640897631645, -0.004373508505523205, ...
laion/CLIP-ViT-bigG-14-laion2b_s39b_b160k
1,280
37
[ 0.015250771306455135, -0.02735118940472603, -0.013550868257880211, 0.007402605842798948, 0.013235434889793396, 0.002909219590947032, 0.01880788244307041, -0.02393045462667942, 0.023575203493237495, 0.009739289991557598, -0.03364070877432823, -0.042035046964883804, 0.023679424077272415, 0.0...
laion/CLIP-ViT-bigG-14-laion2b_s39b_b160k
1,280
38
[ 0.009891790337860584, -0.029506320133805275, -0.014967714436352253, 0.015486819669604301, -0.0020132693462073803, 0.019207805395126343, 0.016290461644530296, -0.03214818239212036, 0.017455194145441055, 0.006097801029682159, -0.02155284211039543, -0.02922757714986801, 0.012088416144251823, ...
laion/CLIP-ViT-bigG-14-laion2b_s39b_b160k
1,280
39
[ 0.04959122836589813, -0.034225404262542725, -0.0312807634472847, 0.02676331251859665, 0.017365988343954086, -0.003367743920534849, 0.009597627446055412, -0.02465919964015484, 0.030281253159046173, -0.021473471075296402, -0.022487271577119827, -0.04552324116230011, 0.00224133487790823, 0.03...
laion/CLIP-ViT-bigG-14-laion2b_s39b_b160k
1,280
40
[ 0.03140822425484657, 0.009088490158319473, -0.014816813170909882, 0.02273097261786461, -0.014775008894503117, -0.0014003481483086944, 0.01210768148303032, -0.043336477130651474, -0.0027955039404332638, 0.03179843723773956, 0.024479340761899948, -0.09733344614505768, 0.025726687163114548, -...
laion/CLIP-ViT-bigG-14-laion2b_s39b_b160k
1,280
41
[ 0.03152276948094368, 0.006117468234151602, -0.012164602056145668, 0.02220107987523079, -0.009119036607444286, -0.016712071374058723, 0.03424280509352684, -0.039518795907497406, 0.014133649878203869, 0.008205241523683071, 0.006417430471628904, -0.07919608801603317, 0.018440643325448036, -0....
laion/CLIP-ViT-bigG-14-laion2b_s39b_b160k
1,280
42
[0.02044248767197132,-0.014504254795610905,0.011645158752799034,-0.00472006481140852,0.0530826821923(...TRUNCATED)
laion/CLIP-ViT-bigG-14-laion2b_s39b_b160k
1,280
43
[0.020442580804228783,-0.014504284597933292,0.01164509728550911,-0.004720060154795647,0.053082592785(...TRUNCATED)
laion/CLIP-ViT-bigG-14-laion2b_s39b_b160k
1,280
44
[-0.011610254645347595,-0.009151593782007694,0.01903832145035267,0.012820248492062092,0.034949332475(...TRUNCATED)
laion/CLIP-ViT-bigG-14-laion2b_s39b_b160k
1,280
End of preview. Expand in Data Studio

metmuseum/openaccess-embeddings-openclip-vitg14

Image embeddings for every public-domain artwork in metmuseum/openaccess, produced by laion/CLIP-ViT-bigG-14-laion2B-s39B-b160K.

Column Type Notes
objectID int64 Primary key — matches objectID in metmuseum/openaccess
embedding list<float32> L2-normalised, dim = 1280
model string Source model id
dim int32 Embedding dimension

Image bytes are not stored here; join against the main dataset to recover them.

Embedding spec: dim=1280, expected image size=224px.

Joining with the main dataset

from datasets import load_dataset

meta = load_dataset("metmuseum/openaccess", split="train")
emb  = load_dataset("metmuseum/openaccess-embeddings-openclip-vitg14", split="train")

# Build an objectID -> embedding lookup, then attach to the metadata rows.
lookup = {r["objectID"]: r["embedding"] for r in emb}
joined = meta.map(lambda r: {"embedding": lookup.get(r["objectID"])})
print(joined[0].keys())

Nearest-neighbour example

import numpy as np
from datasets import load_dataset

emb = load_dataset("metmuseum/openaccess-embeddings-openclip-vitg14", split="train")
ids = np.array(emb["objectID"])
mat = np.array(emb["embedding"], dtype=np.float32)  # already L2-normalised

query = mat[0]
scores = mat @ query
top = np.argsort(-scores)[:10]
print(list(zip(ids[top].tolist(), scores[top].tolist())))

Generated by et-openaccess-embeddings.

Downloads last month
26