Yiqing Huang
Yiqing Huang
How can I provide a custom input image to the pretrained model and get output captions for the same?
You need to use a pre-trained faster R-CNN to extract the object features and put these features into my model. The details of the pre-trained faster R-CNN can be found...
Did you generate the cached tokens before running scst? You can refer to more details about scst in Ruotian Luo's webpage.
My modifications are adding two loss functions but not changing the original SCST loss. Are you using python 3.X? The python version may matter. You can try computing the Cider-D...
It depends on the hardwares like GPU, cpu and RAM, etc. Personally, I suggest that it's better not use only 1 gpu to train captioning models. It will take several...
It seems that there are no files in this drive. Could you please share the data again?
> @RubickH It was the link authors provided in their ReadMe. They might have just removed the data from gdrive. I guess you could try contacting the authors for any...