metadata
license: apache-2.0
configs:
- config_name: default
data_files:
- split: all.AA.train
path: data/all.AA.train-*
- split: all.AA.val
path: data/all.AA.val-*
- split: all.AA.test
path: data/all.AA.test-*
- split: all.RNA.train
path: data/all.RNA.train-*
- split: all.RNA.val
path: data/all.RNA.val-*
- split: all.RNA.test
path: data/all.RNA.test-*
- split: all.DNA.train
path: data/all.DNA.train-*
- split: all.DNA.val
path: data/all.DNA.val-*
- split: all.DNA.test
path: data/all.DNA.test-*
dataset_info:
features:
- name: Sequence
dtype: string
- name: Value
dtype: float64
- name: Domain
sequence: string
splits:
- name: all.AA.train
num_bytes: 48674548
num_examples: 256429
- name: all.AA.val
num_bytes: 9602334
num_examples: 45208
- name: all.AA.test
num_bytes: 20482098
num_examples: 108232
- name: all.RNA.train
num_bytes: 384788975
num_examples: 729302
- name: all.RNA.val
num_bytes: 77418426
num_examples: 148014
- name: all.RNA.test
num_bytes: 78415054
num_examples: 149884
- name: all.DNA.train
num_bytes: 384788975
num_examples: 729302
- name: all.DNA.val
num_bytes: 77418426
num_examples: 148014
- name: all.DNA.test
num_bytes: 78415054
num_examples: 149884
download_size: 171460471
dataset_size: 1160003890
Introduction
This is the official dataset repo of TadABench-1M.
The whole dataset is in the all.xxx.xxx split.
- The protein version of TadABench-1M:
all.AA.train,all.AA.test,all.AA.val. - The DNA version of TadABench-1M:
all.DNA.train,all.DNA.test,all.DNA.val. - The RNA version of TadABench-1M:
all.RNA.train,all.RNA.test,all.RNA.val.
Usage
Check all the splits.
from datasets import load_dataset
# Login using e.g. `huggingface-cli login` to access this dataset
ds = load_dataset("JinGao/TadABench-1M", split=None, streaming=True)
Load the part you want, e.g., the test set of the whole protein version.
from datasets import load_dataset
from itertools import islice
ds = load_dataset("JinGao/TadABench-1M", split="all.AA.test", streaming=True)
first_10 = list(islice(ds, 10))
print(first_10)