Using Vatex dataset to train my own chinese data

[orangejhang]

I want to use the Vatex dataset to train my own chinese data. I count the len of chinese word according by the unicode format. I also use json parser to check each line of format, but it appear is correct. it's successed to complete one epoch, but I got an error at second epochs.

self.image_keys[img_idx]: datasets/VATE2/raw_videos/val_all/A19_2.mp4
train(args, train_dataloader, val_dataloader, vl_transformer, tokenizer, training_saver, optimizer, scheduler)
File "src/tasks/run_caption_VidSwinBert.py", line 146, in train
for iteration, (img_keys, batch, meta_data) in enumerate(train_dataloader):
File "/opt/conda/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 517, in next
data = self._next_data()
File "/opt/conda/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 1179, in _next_data
return self._process_data(data)
File "/opt/conda/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 1225, in _process_data
data.reraise()
File "/opt/conda/lib/python3.8/site-packages/torch/_utils.py", line 429, in reraise
raise self.exc_type(msg)
TypeError: init() missing 2 required positional arguments: 'doc' and 'pos'

[tiesanguaixia]

Hi! May I ask how to download the raw videos of VATEX?

[markllai]

Hi! May I ask how to download the raw videos of VATEX?

use youtube-dl or yt-dlp

[Nishanta45]

how to download raw video of vatex dataset of 10 second including start time and end time