No module named 'refTools'

[PeterDykas]

Hi I am trying to reproduce the results from inside a docker container. After installing the dependencies I hit the following error:

  File "Pretrain.py", line 30, in <module>
    from dataset import create_dataset, create_sampler, create_loader
  File "/workspace/dataset/__init__.py", line 6, in <module>
    from dataset.caption_dataset import re_train_dataset, re_eval_dataset, pretrain_dataset
  File "/workspace/dataset/caption_dataset.py", line 12, in <module>
    from dataset.utils import pre_caption
  File "/workspace/dataset/utils.py", line 45, in <module>
    from refTools.evaluation.refEvaluation import RefEvaluation
ModuleNotFoundError: No module named 'refTools'

every time when running:

python -m torch.distributed.launch --nproc_per_node=8 \ --use_env Pretrain.py \ --config ./configs/Pretrain.yaml \ --output_dir output/pretrain

I have tried pip3 install reftools but it does not solve the issue. Have you run into this issue before?

[viyjy]

Hi, thanks for your interest of our work. Let me double check this issue on my machine, will let you know ASAP.

[viyjy]

Hi, please download refTools , then unzip it, then put it under your TCL folder. Please let me know if you might need any further information.

[PeterDykas]

That seems to have fixed the issue thank you. If I want to run on only a subset of the pretraining data is commenting out unused datasets in the config sufficient?

[viyjy]

Yes, just modify train_file in Pretrain.yaml.

[PeterDykas]

Awesome thanks, I see that you listed MSCOCO 2014 as a dataset but I am going to assume it means the MSCOCO 2015 where the captioning competition was.

[viyjy]

Please only download MSCOCO 2014 datasets from these links.
2014 Train images [83K/13GB]
2014 Val images [41K/6GB]
2014 Test images [41K/6GB]