diff --git a/jsonl_data/jsonl_to_json.py b/jsonl_data/jsonl_to_json.py new file mode 100644 index 00000000..eb6917b4 --- /dev/null +++ b/jsonl_data/jsonl_to_json.py @@ -0,0 +1,20 @@ +import json + +def jsonl_to_json(jsonl_file, json_file): + with open(jsonl_file, 'r', encoding='utf-8') as f: + jsonl_data = f.readlines() + # print(jsonl_data) + # print(jsonl_data[0]) + # print(jsonl_data[0][1:-2]) + # print(jsonl_data[14]) + + with open(json_file, 'w', encoding='utf-8') as f: + for line in jsonl_data: + line = line[1:-2] + f.write(' ' + line + ',' + '\n') + f.close() + + +jsonl_file = 'jsonl_data/alpaca_zh.jsonl' +json_file = 'data/alpaca_zh.json' +jsonl_to_json(jsonl_file, json_file)