njucckevin/knowcap

Code for Beyond Generic: Enhancing Image Captioning with Real-World Knowledge using Vision-Language Pre-Training Model

knowcap's Issues

What is the version of transformers

There has been a problem when i try to reproduce the results, i cannot import GitConfig, GitVisionConfig from transformers.models.git.configuration_git. I cannot import OFATokenizer from transformers.models.ofa.tokenization_ofa in test_knowcao.py too. I have tried with 4.30.2 and 4.35.2. I wanna know that which version can be used to run the code. I would appreciate for your answer.

Unable to install transformers4.28.0.dev0

I've tried a lot of methods but can't install the 4.28.0.dev0 version of transformers, can you provide an installation method, thank you very much

'str' object has no attribute 'getImgIds'

Traceback (most recent call last):
File "train_multitask.py", line 228, in
pycoco_results = eval_pycoco(config, gen_pycoco_path, 'val')
File "/home/tuxiangsuo1502/KnowCap/utils/eval.py", line 80, in eval_pycoco
cocoEval = COCOEvalCap('diy', 'diy')
File "/home/tuxiangsuo1502/miniconda3/envs/hjl/lib/python3.7/site-packages/pycocoevalcap/eval.py", line 17, in init
self.params = {'image_id': coco.getImgIds()}
AttributeError: 'str' object has no attribute 'getImgIds
Do you have any clue when I encountered the above problems while starting training from scratch Thank you in advance for your reply