metadata
dataset_info:
features:
- name: text
sequence: string
- name: labels
sequence:
class_label:
names:
'0': O
'1': B-NS
'2': M-NS
'3': E-NS
'4': S-NS
'5': B-NT
'6': M-NT
'7': E-NT
'8': S-NT
'9': B-NR
'10': M-NR
'11': E-NR
'12': S-NR
splits:
- name: train
num_bytes: 32917977
num_examples: 46364
- name: validation
num_bytes: 2623860
num_examples: 4365
- name: test
num_bytes: 2623860
num_examples: 4365
download_size: 4762958
dataset_size: 38165697
How to loading dataset?
from datasets import load_dataset
datasets = load_dataset("minskiter/msra_dev",save_infos=True)
train,test = datasets['train'],datasets['test']
# convert label to str
print(train.features['labels'].feature.int2str(0))
Force update
from datasets import load_dataset
datasets = load_dataset("minskiter/msra_dev", download_mode="force_redownload")
Fit your train
def transform(example):
# edit example here
return example
for key in datasets:
datasets[key] = datasets.map(transform)