tencentarc / mcq Goto Github PK

Hello, wonderful project!. Here I wonder how to finetune the pre-trained models on downstream video-text retrieval datasets like MSR-VTT, LSMDC, and MSVD? I notice that the script for zero-shot retrieval has been provided, but there is no script about how to finetune on retrieval datasets.

我的意思是CLIP-initialized model 的MCQ模型代码，特别是BridgeFormer与VideoFormer和TextFormer的交互部分。

hello, I want to know whether the result in the paper comes from sliding_window_stride=12 or default=-1？

I want to know if there is a regression head for MVM during the MILES pretraining phase.

I have read your paper MCQ and really appreciate it. So do you have a plan to release the code?

Thanks to your great work!
I want to extract noun phrases and verbs on my own dataset, could you please tell me what tool you used to extract it?

As mentioned in table 4, there are 3 different test split. How are the specific test sets selected and how many are there? Also for the table 5, what is the training data and what is the test data

Given a video I want to do captioning, or as you sugest answer questioning? Is it something possible?

Thx !

论文表2(a)中msvd检索结果引用了Frozen的，zero-shot为33.7，fine-tuning为45.6。请问这个结果是复现得到还是原论文中的？我在原Frozen论文中只看到了一个33.7的R@1，应该是fine-tuning的结果。

请问您能否共享一下去除三元组后的数据集

Hi, I want to know how to extract the phrase in the paper? I saw the issue that mentioned extracting the noun phrases, but it did not consistent what presented in the paper. For example, how to extract "an old woman" rather than "woman"?

Is there any scripts that I can used for extracting the phrases?

Hi,
I'm wondering why you add three [MASK] in answers. I have seen your reply in #7, but I still don't know why the number of [MASK] and whether it is important.
Any reply will be helpful!
Thank you for your good job again.

Hello, thank you for the code of MCQ! We utilize the released weights and follow the data settings, trying to reproduce MSRVTT ZS results. But our result(R@1) is about four points lower than the reported result in the paper. Is there any place we need to pay attention to? Thank you.

Hi, is there any script or config that use CLIP as the initialization?