Dataset Viewer
Auto-converted to Parquet Duplicate
id
int64
0
5k
image
imagewidth (px)
200
640
image_id
stringlengths
2
6
filename
stringlengths
16
16
captions
listlengths
5
7
caption
stringlengths
30
138
image_emb
list
text_emb
list
0
38
000000179765.jpg
[ "A black Honda motorcycle parked in front of a garage.", "A Honda motorcycle parked in a grass driveway", "A black Honda motorcycle with a dark burgundy seat.", "Ma motorcycle parked on the gravel in front of a garage", "A motorcycle with its brake extended standing outside" ]
A black Honda motorcycle parked in front of a garage.
[ -0.0203704833984375, -0.1126708984375, 0.002735137939453125, -0.005603790283203125, -0.03448486328125, 0.0124053955078125, -0.004077911376953125, -0.0012073516845703125, 0.037933349609375, 0.066650390625, -0.043731689453125, 0.00951385498046875, -0.02099609375, -0.030609130859375, 0.0398...
[ 0.027130126953125, -0.07635498046875, 0.06671142578125, -0.09283447265625, 0.02606201171875, -0.0034084320068359375, -0.0190582275390625, -0.014678955078125, 0.02374267578125, 0.0631103515625, -0.048126220703125, 0.019439697265625, 0.031829833984375, -0.020477294921875, 0.094970703125, ...
1
401
000000190236.jpg
[ "An office cubicle with four different types of computers.", "The home office space seems to be very cluttered.", "an office with desk computer and chair and laptop.", "Office setting with a lot of computer screens.", "A desk and chair in an office cubicle." ]
An office cubicle with four different types of computers.
[ 0.00457000732421875, 0.08441162109375, 0.016754150390625, -0.0211639404296875, 0.0007762908935546875, 0.0067901611328125, -0.029998779296875, -0.0294342041015625, 0.01898193359375, 0.0193328857421875, 0.0550537109375, -0.011199951171875, -0.0119171142578125, 0.004352569580078125, -0.0268...
[ 0.0350341796875, -0.0139007568359375, 0.042724609375, 0.0099334716796875, 0.0177001953125, 0.0044403076171875, -0.043487548828125, -0.005611419677734375, -0.004009246826171875, -0.017608642578125, 0.006256103515625, -0.0306549072265625, 0.0179901123046875, -0.03155517578125, 0.0202484130...
2
441
000000331352.jpg
[ "A small closed toilet in a cramped space.", "A tan toilet and sink combination in a small room.", "This is an advanced toilet with a sink and control panel.", "A close-up picture of a toilet with a fountain.", "Off white toilet with a faucet and controls. " ]
A small closed toilet in a cramped space.
[ 0.0028896331787109375, 0.0025348663330078125, -0.031158447265625, 0.06689453125, 0.02593994140625, 0.01355743408203125, -0.01308441162109375, 0.03558349609375, 0.00003266334533691406, 0.00801849365234375, -0.06903076171875, -0.0325927734375, -0.0029850006103515625, 0.00928497314453125, -...
[ 0.0703125, -0.0019216537475585938, 0.052337646484375, 0.04461669921875, 0.04278564453125, 0.0263671875, -0.01032257080078125, 0.01143646240234375, -0.0228271484375, 0.0252838134765625, -0.0255126953125, -0.011871337890625, 0.00921630859375, 0.015716552734375, -0.047943115234375, -0.001...
3
447
000000517069.jpg
[ "Two women waiting at a bench next to a street.", "A woman sitting on a bench and a woman standing waiting for the bus.", "A woman sitting on a bench in the middle of the city", "A woman sitting on a bench and a woman standing behind the bench at a bus stop", "A woman and another woman waiting at a stop." ]
Two women waiting at a bench next to a street.
[ -0.0079193115234375, 0.11749267578125, -0.1387939453125, 0.08599853515625, 0.002613067626953125, -0.0161895751953125, 0.044677734375, -0.0435791015625, 0.001735687255859375, 0.002422332763671875, 0.039337158203125, 0.0115203857421875, -0.00537109375, 0.006824493408203125, 0.0244598388671...
[ 0.0426025390625, 0.00830841064453125, 0.0989990234375, 0.125244140625, 0.004150390625, 0.0211639404296875, -0.04425048828125, -0.0758056640625, 0.0137481689453125, 0.0005955696105957031, -0.01203155517578125, 0.0242767333984375, -0.01020050048828125, -0.0264739990234375, -0.0136032104492...
4
856
000000182417.jpg
[ "A beautiful dessert waiting to be shared by two people", "There is a piece of cake on a plate with decorations on it.", "Creamy cheesecake dessert with whip cream and caramel.", "An extravagant dessert on a plate overlooking the water.", "This is a picture of an extremely fancy desert." ]
A beautiful dessert waiting to be shared by two people
[ -0.06353759765625, 0.10516357421875, -0.064697265625, 0.080810546875, -0.0205230712890625, -0.02203369140625, -0.007602691650390625, -0.006923675537109375, 0.0185546875, 0.006534576416015625, -0.04888916015625, 0.017730712890625, -0.01348876953125, 0.0684814453125, 0.01904296875, -0.01...
[ -0.0136871337890625, 0.06939697265625, 0.0268707275390625, 0.0819091796875, -0.011077880859375, -0.0106048583984375, -0.01485443115234375, -0.01496124267578125, -0.0166778564453125, 0.00034427642822265625, -0.0498046875, -0.0074615478515625, -0.0012845993041992188, -0.00215911865234375, ...
5
1158
000000046378.jpg
[ "A cat eating a bird it has caught.", "A white cat caught a bird outside on a patio.", "Grey house cat devours a song bird on a door step", "A long haired cat eating a dead bird.", "A cat eating a dead bird on the ground." ]
A cat eating a bird it has caught.
[ -0.00046634674072265625, 0.1083984375, -0.09954833984375, 0.04229736328125, -0.060302734375, -0.0187835693359375, 0.019561767578125, 0.0007672309875488281, 0.025482177734375, -0.047149658203125, -0.03887939453125, 0.0172882080078125, 0.0252685546875, -0.03082275390625, 0.030364990234375,...
[ 0.01465606689453125, 0.01175689697265625, 0.0019407272338867188, 0.042327880859375, -0.03143310546875, 0.0234222412109375, 0.012847900390625, -0.0175323486328125, -0.022308349609375, -0.060028076171875, -0.0670166015625, -0.0138702392578125, 0.0014486312866210938, 0.0312042236328125, 0.0...
6
1437
000000093437.jpg
[ "A shot of an elderly man inside a kitchen.", "An old man is wearing an odd hat", "An older man is wearing a funny hat in his dining room.", "A man in a jacket and hat looks at the camera.", "An old man standing in a kitchen posing for a picture." ]
A shot of an elderly man inside a kitchen.
[ -0.038970947265625, 0.09771728515625, -0.03302001953125, 0.0272064208984375, -0.0265045166015625, -0.009979248046875, -0.0008263587951660156, -0.00516510009765625, 0.035247802734375, 0.003246307373046875, 0.024688720703125, -0.050140380859375, -0.010467529296875, -0.0028476715087890625, ...
[ -0.0240325927734375, -0.0021419525146484375, 0.0328369140625, 0.006381988525390625, 0.00475311279296875, -0.00518035888671875, -0.044464111328125, -0.006938934326171875, -0.033233642578125, -0.013427734375, -0.0231781005859375, 0.00246429443359375, 0.01120758056640625, 0.030487060546875, ...
7
1536
000000172330.jpg
[ "A cat in between two cars in a parking lot.", "A cat stands between two parked cars on a grassy sidewalk. ", "A cat at attention between two parked cars.", "A grey and white cat watches from between parked cars.", "A grey and white cat standing in the grass in a parking lot. " ]
A cat in between two cars in a parking lot.
[ 0.0306854248046875, 0.083740234375, -0.097900390625, 0.049407958984375, -0.0026340484619140625, 0.05108642578125, 0.007602691650390625, -0.033843994140625, 0.005519866943359375, 0.01514434814453125, 0.00559234619140625, -0.03125, 0.023681640625, 0.0160980224609375, -0.02557373046875, 0...
[ -0.0257568359375, -0.04046630859375, 0.0266876220703125, -0.0168304443359375, 0.0247039794921875, 0.0623779296875, 0.0271759033203125, -0.004077911376953125, -0.017730712890625, -0.005039215087890625, -0.05010986328125, 0.015594482421875, 0.021240234375, -0.037017822265625, 0.03543090820...
8
1638
000000472678.jpg
[ "An office cubicle with multiple computers in it", "An office desk with two flat panel monitors.", "An office desk with two computer screens, books diagrams and a phone on it.", "Two computer monitors are placed beside each other on a desk.", "A desk with two monitors depicting security cameras." ]
An office cubicle with multiple computers in it
[ -0.018096923828125, 0.0268402099609375, 0.0006690025329589844, -0.033782958984375, -0.00661468505859375, 0.0008416175842285156, -0.029327392578125, -0.048248291015625, -0.0250396728515625, -0.028961181640625, 0.0259552001953125, 0.011993408203125, 0.018768310546875, 0.04290771484375, -0....
[ 0.0640869140625, -0.01226043701171875, 0.01995849609375, -0.0189666748046875, 0.029876708984375, -0.0081939697265625, -0.041595458984375, -0.007190704345703125, -0.0050201416015625, -0.00835418701171875, 0.038726806640625, -0.01215362548828125, 0.0162506103515625, -0.037994384765625, 0.0...
9
1879
000000314251.jpg
[ "A parade of motorcycles is going through a group of tall trees.", "A group of motorcyclists drive down a tree lined street.", "A group of motorcycles down a long street filled with trees on either side.", "A group of people riding mopeds through a park.", "A group of scooters rides down a street" ]
A parade of motorcycles is going through a group of tall trees.
[ 0.0419921875, 0.0908203125, -0.126953125, -0.013153076171875, -0.0037555694580078125, 0.0013942718505859375, 0.004970550537109375, -0.038604736328125, 0.00722503662109375, 0.01038360595703125, -0.029266357421875, -0.0227203369140625, -0.0111541748046875, 0.036376953125, 0.008827209472656...
[ 0.041290283203125, -0.012481689453125, -0.04840087890625, -0.012664794921875, 0.00756072998046875, -0.034088134765625, -0.0257568359375, 0.02130126953125, -0.003276824951171875, -0.058929443359375, -0.0307159423828125, -0.01271820068359375, 0.01168060302734375, -0.0216522216796875, 0.051...
End of preview. Expand in Data Studio

COCO Captions 2017 (Lance Format)

Lance-formatted version of the COCO Captions 2017 corpus, redistributed via lmms-lab/COCO-Caption2017. Each row is one image with 5–7 human-written captions, CLIP image embedding, and CLIP text embedding of the canonical caption — all stored inline.

Splits

Split Rows
val.lance 5,000 (canonical COCO 2017 val set)
test.lance 40,700

The 2017 train split (118 k images, ~18 GB of source JPEGs) is intentionally not bundled here because the lmms-lab/COCO-Caption2017 redistribution does not include it. To extend with train, run coco_captions_2017/dataprep.py against your local COCO 2017 train mirror.

Schema

Column Type Notes
id int64 Row index within split
image large_binary Inline JPEG bytes
image_id string COCO image id
filename string Original filename (e.g. 000000179765.jpg)
captions list<string> All 5–7 captions
caption string First caption — used as canonical text for FTS
image_emb fixed_size_list<float32, 512> CLIP image embedding (cosine-normalized)
text_emb fixed_size_list<float32, 512> CLIP text embedding of the canonical caption

Pre-built indices

  • IVF_PQ on image_emb and text_embmetric=cosine
  • INVERTED on caption
  • BTREE on image_id

Quick start

import lance

ds = lance.dataset("hf://datasets/lance-format/coco-captions-2017-lance/data/val.lance")
print(ds.count_rows(), ds.schema.names)
print(ds.list_indices())

Load with LanceDB

These tables can also be consumed by LanceDB, the multimodal lakehouse and embedded search library built on top of Lance, for simplified vector search and other queries.

import lancedb

db = lancedb.connect("hf://datasets/lance-format/coco-captions-2017-lance/data")
tbl = db.open_table("val")
print(f"LanceDB table opened with {len(tbl)} image-caption pairs")

Tip — for production use, download locally first.

hf download lance-format/coco-captions-2017-lance --repo-type dataset --local-dir ./coco-captions-2017-lance

Vector search examples

Cross-modal text→image:

import lance, open_clip, pyarrow as pa, torch

model, _, _ = open_clip.create_model_and_transforms("ViT-B-32", pretrained="laion2b_s34b_b79k")
tokenizer = open_clip.get_tokenizer("ViT-B-32")
model = model.eval().cuda().half()
with torch.no_grad():
    q = model.encode_text(tokenizer(["a giraffe eating leaves"]).cuda())
    q = (q / q.norm(dim=-1, keepdim=True)).float().cpu().numpy()[0]

ds = lance.dataset("hf://datasets/lance-format/coco-captions-2017-lance/data/val.lance")
emb_field = ds.schema.field("image_emb")
hits = ds.scanner(
    nearest={"column": "image_emb", "q": pa.array([q.tolist()], type=emb_field.type)[0], "k": 10},
    columns=["image_id", "caption"],
).to_table().to_pylist()

LanceDB cross-modal text→image search

import lancedb, open_clip, torch

model, _, _ = open_clip.create_model_and_transforms("ViT-B-32", pretrained="laion2b_s34b_b79k")
tokenizer = open_clip.get_tokenizer("ViT-B-32")
model = model.eval().cuda().half()
with torch.no_grad():
    q = model.encode_text(tokenizer(["a giraffe eating leaves"]).cuda())
    q = (q / q.norm(dim=-1, keepdim=True)).float().cpu().numpy()[0]

db = lancedb.connect("hf://datasets/lance-format/coco-captions-2017-lance/data")
tbl = db.open_table("val")

results = (
    tbl.search(q.tolist(), vector_column_name="image_emb")
    .metric("cosine")
    .select(["image_id", "caption"])
    .limit(10)
    .to_list()
)

Full-text search:

ds = lance.dataset("hf://datasets/lance-format/coco-captions-2017-lance/data/val.lance")
hits = ds.scanner(
    full_text_query="surfer riding a wave",
    columns=["image_id", "caption"],
    limit=10,
).to_table().to_pylist()

LanceDB full-text search

import lancedb

db = lancedb.connect("hf://datasets/lance-format/coco-captions-2017-lance/data")
tbl = db.open_table("val")

results = (
    tbl.search("surfer riding a wave")
    .select(["image_id", "caption"])
    .limit(10)
    .to_list()
)

Why Lance?

  • One dataset carries images + image embeddings + text embeddings + indices — no sidecar files.
  • On-disk vector and full-text indices live next to the data, so search works on local copies and on the Hub.
  • Schema evolution: add columns (new captions, alternate embeddings, model predictions) without rewriting the data.

Source & license

Converted from lmms-lab/COCO-Caption2017. Original COCO 2017 annotations are released under CC BY 4.0; the underlying images are subject to Flickr terms of service. Please review the COCO Terms of Use before redistribution.

Citation

@inproceedings{lin2014microsoft,
  title={Microsoft COCO: Common objects in context},
  author={Lin, Tsung-Yi and Maire, Michael and Belongie, Serge and Hays, James and Perona, Pietro and Ramanan, Deva and Doll{\'a}r, Piotr and Zitnick, C Lawrence},
  booktitle={European Conference on Computer Vision (ECCV)},
  year={2014},
}
Downloads last month
52