| import json | |
| import random | |
| #json_path = "/data/work-gcp-europe-west4-a/yuqian_fu/Ego/data_segswap/egoexo_val_framelevel_all.json" | |
| json_path = "/data/work-gcp-europe-west4-a/yuqian_fu/datasets/HANDAL/handal_dataset_mugs/handal_train.json" | |
| output_path = "/data/work-gcp-europe-west4-a/yuqian_fu/datasets/HANDAL/handal_dataset_mugs/handal_test100.json" | |
| #random.seed(42) | |
| # 读取原始的完整数据 | |
| with open(json_path, 'r') as f: | |
| data_full = json.load(f) | |
| # # 计算三分之一的数据量 | |
| # third_size = len(data_full) // 3 | |
| # # 随机选择三分之一的数据 | |
| # # data_smallsize = random.sample(data_full, third_size) | |
| # # 按照原顺序，每间隔三个数据存储一个 | |
| #data_smallsize = data_full[:100] | |
| data_smallsize = random.sample(data_full, 100) | |
| # 将缩小后的数据保存到新的json文件中 | |
| with open(output_path, 'w') as f: | |
| json.dump(data_smallsize, f) | |
| print(f'Saved {len(data_smallsize)} samples to smallsize_version') |