CoolFace
Datasetpublic

dn-gh/Tahoe-x1-CRC-embeddings

Tahoe-x1 3B — CRC slice Cell embeddings from tahoebio/Tahoe-x1-embeddings (Tx1-3B on Tahoe-100M), restricted to CRC cell lines. This is a filter of the published embeddings, not a new model run. Source license: Apache-2.0. Cells: 21,068,133 line CVCL n cells SW480 CVCL_0546 6,040,371 LoVo CVCL_0399 3,013,246 RKO CVCL_0504 2,182,314 HT-29 CVCL_0320 2,171,036 SW1417 CVCL_1717 1,921,742 LS 180 CVCL_0397 1,828,299 HCT15 CVCL_0292 1,500,121 COLO 205 CVCL_0218… See the full description on the dataset page: https://huggingface.co/datasets/dn-gh/Tahoe-x1-CRC-embeddings.

sourceHugging Faceapache-2.0updated 16d agoView on Hugging Face
0likes184downloads
Dataset Card

Tahoe-x1 3B — CRC slice

Cell embeddings from `tahoebio/Tahoe-x1-embeddings` (Tx1-3B on Tahoe-100M), restricted to CRC cell lines.

This is a filter of the published embeddings, not a new model run. Source license: Apache-2.0.

Cells: 21,068,133

lineCVCLn cells
SW480CVCL_05466,040,371
LoVoCVCL_03993,013,246
RKOCVCL_05042,182,314
HT-29CVCL_03202,171,036
SW1417CVCL_17171,921,742
LS 180CVCL_03971,828,299
HCT15CVCL_02921,500,121
COLO 205CVCL_02181,456,805
SW48CVCL_1724954,199

Columns

Same as the source dataset:

columnmeaning
drugcompound name
sampleTahoe sample id
cell_lineCellosaurus id
BARCODE_SUB_LIB_IDcell barcode
mosaicfm-3b-prod-cont-MFMv2Tx1-3B embedding, 2560-d float32

Load

python
from datasets import load_dataset
ds = load_dataset("dn-gh/Tahoe-x1-CRC-embeddings", split="train", streaming=True)

Do not confuse this with the Tahoe-100M expression matrices (h5ad). This repo is embeddings only.