Ice

HI FengNian, yes modify it from N frames to 10 frames. yes after running then perform optical flow extraction. Sorry I have not searched my files, will do so promptly...

@ILoveXuXin Thanks for the effort and kind sharing. Hope it helps for the rest. :)

I have, but thinking how to send to you because 20gb in size.

Because each different fold has its own weight (cross validation), so in total there are 26 spatial and 26 temporal weights.

weights for dige.rar -- *Journey of thousand miles begins with a single step.* On Fri, 24 Aug 2018 at 15:01, Dige Ai wrote: > oh, sorry I have forgotten this....

input variable should be (237, 9, 4096). `features` should be (4096), take the 2nd last layer of vgg_16(a.k.a. one layer before softmax, the `Dense(4096)`). LSTM should be the module that...

I see, maybe was a typo. :P Tested for Spatial Module. Thanks for the contribution!!! How about your side? does it work for full training?

I also use theano. tensorflow keras overflows a 1080TI as well. not recommended. batch size around 20-30 should be fine. if still doesn't work, try reducing the lstm's recurrent units...

Ya if you changed to MF1, it would drop a little bit. Sorry because I am not too sure about the problem, there are too much factors that could affect...

Ya a bit of confusion. But the way I calculated it is summing all tp, fp and fn. U can find it in evaluationmatrix.py On Wed, 15 Jan 2020, 17:11...