Image Captionng

gistfile1.txt

with open('open_images_train_captions.jsonl', 'r') as json_file:
    json_list = json_file.read().split('\n')
np.random.shuffle(json_list)
data = []
N = 100000
for ix, json_str in Tqdm(enumerate(json_list), N):
    if ix == N: break
    try:
      result = json.loads(json_str)
      x = pd.DataFrame.from_dict(result, orient='index').T
      data.append(x)
    except:
      pass

update