1-2对联数据转化成jsonl格式

使用python工具

通过csv方式读取对联数据

上联读取和下联读取分开进行

合并上下联数据

转化成jsonl格式

# 导入所需的库 import pandas as pd import jsonlines import csv # 读取train.csv文件内容为DataFrame jsonl_in = [] jsonl_out = [] x=[] path = 'duilian\\1fixed_couplets_in-1.csv' with open(path, 'r', encoding='utf-8') as f: reader = csv.reader(f) for row in reader: jsonl_in.append(row[0].replace(' ', '')) if len(jsonl_in)>4000: break # print(jsonl_in) path = 'duilian\\1fixed_couplets_out-1.csv' with open(path, 'r', encoding='utf-8') as f1: reader_output = csv.reader(f1) for row1 in reader_output: jsonl_out.append(row1[0].replace(' ', '')) if len(jsonl_out)>4000: break for i in range(4000): value = {} value['instruction'] = '请根据上联，回复下联，要求对仗工整' value['input']=jsonl_in[i] value['output']=jsonl_out[i] x.append(value) with jsonlines.open('duilian.jsonl', mode='w') as writer: writer.write_all(x)

# 导入所需的库
import pandas as pd
import jsonlines
import csv
# 读取train.csv文件内容为DataFrame

jsonl_in = []
jsonl_out = []
x=[]
path = 'duilian\\1fixed_couplets_in-1.csv'
with open(path, 'r', encoding='utf-8') as f:
    reader = csv.reader(f)
    for row in reader:

        jsonl_in.append(row[0].replace(' ', ''))
        if len(jsonl_in)>4000:
            break
    # print(jsonl_in)
path = 'duilian\\1fixed_couplets_out-1.csv'
with open(path, 'r', encoding='utf-8') as f1:
    reader_output = csv.reader(f1)
    for row1 in reader_output:
        jsonl_out.append(row1[0].replace(' ', ''))
        if len(jsonl_out)>4000:
            break

for i in range(4000):
    value = {}
    value['instruction'] = '请根据上联，回复下联，要求对仗工整'
    value['input']=jsonl_in[i]
    value['output']=jsonl_out[i]

    x.append(value)
with jsonlines.open('duilian.jsonl', mode='w') as writer:
    writer.write_all(x)