Cathy

Problems when memcnn is used with data parallel

1

Hello~ I am wondering if memcnn works for data parallel. When I use it with data parallell for multi-GPU running, all layers have requires_grad=False. * MemCNN version: 1.5.0 * PyTorch...

Hi, @sunnyxiaohu ! R-C3D achieves a very good performance on the Thumos14 dataset. I can reach 0.4175 @ IoU 0.5 using your implementation. However, it does not perform as well...

What learning rates are you using to train the model? I use the default hyper-parameters specified in the provided code and 4 GPUs to train the C3D model on THUMOS14....

Hi! In the paper, it mentions that the query is an ROI, and key and values are from the clip around an ROI. But in this implementation, it seems that...

Why perform rendering transformer by the entire feature map?

1

May I ask why you perform rending transformer by the entire feature map instead of pixels? Does it work well if you do rendering transformer the same way for grounding...

1. There is no `data_file` in the project directory. 2. Could you please provide the link to the dataset you use as well as the annotations? The datapaths in the...

Hi, Rohit! May I ask if I can replace [the model definition using the package 'transformers'](https://github.com/facebookresearch/AVT/blob/b082c99a6b3104780237022274db74f5c3124cc5/models/future_prediction.py#L89) with any other package or code? That would solve my following issue. The 'transformers'...

GPU type for training

1

Hi! Thanks for the great work! May I ask what GPU type did you use for training? Did you also use gtx1080ti the same as inference?