Alexander Werthmann
Alexander Werthmann
You could just add another (trading) day to the time interval by modifying either the start or end date. Then it should fit the shape.
Nice work @BCJuan ! I would like to implement your solution as well (I train with my own dataset too). Could you please show me your code where call the...
@BCJuan Thank you, I found the lines in evaluation and it should be easy to get this working. Did you call the `sess.run(redcued_loss_val)` in a loop running _num_val_images/batch_size_ times in...
If the majority of your groundtruth is class 0 this can happen. I had a similiar problem with my own dataset and solved it by applying class weights to the...
@chiyuan-goh Yes you are right. Bigger weights should be applied to rare features of course. For weight calculation I calculated so called median-frequency weights introduced by Eigen et al. in...
also discussed in this issue. It is still a problem https://github.com/TelegramMessenger/Telegram-iOS/issues/1100
I noticed this too and it kills usability of my bot...
still an issue. tested it on different android phones and it doesnt seem to compress.
So the build was successful ? According to the ReadMe it needs to be built via cmake or bazel . Did you follow the links provided , built it and...
I also noticed the inference increasing significantly making live detection with more than super primitive model very ugly. Do you remember which version started the inference to increase @yingeo ?