LiangXu123

I am curious too,since there is a PR about Integration with vot-toolkit,but it's still remain open,and I try it,but not success, runing in MATLAB result something like this: Verifying native...

it's really does not matter,since all models are trained and tested on standard UCF101 split1 train/test, so you can prepare your own train/test list as the paper says instead of...

@b-safwat I believe the big margin in validation-set accuracy during training phase and test phase is due to the different data normalization , or due to the Dropout layer, since...

hello, when you call saver.restore(sess, model_filename) ),the model_filename parameter can be assigned by: 'model_filename=tf.trainable_variables()' for all trainable parameters in the checkpoint, in your case,: 'model_filename= [v for v in tf.trainable_variables()...

checkout https://github.com/hx173149/C3D-tensorflow/blob/master/C3D-tensorflow-1.0/Random_clip_valid.py, my bad to ignore that model_filename is a path. when you define the saver by : saver = tf.train.Saver() you can also do it by assign variables name:...

I believe it is RGB, since we use PIL.Image to load image and we did not tranpose it to BGR manually.

> try this: > > https://github.com/JonasSchult/Mask3D/blob/5c5eabed4fb415c9de892404963c52e7ec02b68c/models/mask3d.py#L56 > > change to: `self.backbone =config.backbone` you saved my day!

@Feynman27 Hello man,it seems that we two are doing the same thing now，i am still working on training DT in pytorch，and i use correlation layer comes from nvidia flownet2 too....

my result : R-FCN one frame baseline+finetune on DT with 1e-4 lr： 69.7% is got by using alternating sample VID or DET in each iteration,but i did not know how...

and for now，i am trying to use alternating sample from VID or DET in each iteration everywhere，not only in finetune the DT from RFCN as the paper says