STTran
STTran copied to clipboard
yrcong
Spatial-Temporal Transformer for Dynamic Scene Graph Generation, ICCV2021
I.m trying to setup sttran from scratch and I'm getting errors while setting up fasterRCNN package. The command I run is : `STTran/fasterRCNN/lib: python setup.py install` ``` fatal error: THC/THC.h:...
Scene Graph prediction on custom video without downloading any dataset and using pre trained models
I actually just want to use pretrained models files and detect scene graphs in a custom video/images, instead of downloading AG dataset or training right now. But I am not...
Hello, Thanks a lot your nice work. In the paper, you mentioned to evaluate on 25 classes, however, in your dataloader you have "Other Relationships" as a contact class which...
Hi, Can you share the code of computing the average precision of predicates? Thank you! 😊
Is there any suggestion on how to run the model on customized input videos? Thank you!
@yrcong What train-val-test split is used with Action Genome dataset? From the codebase, it seems like the same test set was used as validation set during training (in train.py) and...
In the introduction video on youtube, we see the author visualize the recognition of human actions in video frames. What I want to know is, did the author include the...
Why is the shape of global output different during training and evaluation for the same input video? I have passed the same vid input to train.py and test.py, but the...
the dataset [Action Genome](https://www.actiongenome.org/#download) is cannot be opened
