hanban

Hi, may I ask what is the correct way of calculating bit/dim ( for different n-bits images) from NLL ?

> I think it is easy to run on batches with small modifications, if you are familiar with pytorch. But the model requires cropped images. Any suggestions on batch aligned...

> I did a batched version for a single box. This is intended to be used in videos where there is a single person whose face remains inside the box...

Errors pop after `brew install libpng`, it seems to be m1 version of brew 's problem? libpng installed properly. ``` $ pkg-config --list-all libde265 libde265 - H.265/HEVC video decoder. libpng...

I encountered this issue without using `model.fit()`. I replaced all `K.zeros(K.stack(...))` with `tf.zeros(tf.stack(...))` in the codes and it magically works 😂 ...

Glow has a downsample layer which just throws away a half of data (along channel). In the reverse process, the threw-away part is sampled from a prior. You can modify...

I try same codes on a different machine (almost the same configuration and environment except for Ubuntu OS version I guess), memory is not growing that fast anymore, to some...

As model goes bigger(image size (64,64,3) and num_glow_blocks=4) , memory leak still exist although in a slow way. And SUPER SLOW for graph compiling, takes about 10min to start training....

``` PID USER PRI NI VIRT RES SHR S CPU% MEM% TIME+ Command 34953 root 20 0 126G 102G 1521M S 111. 40.7 1h43:31 python ``` I encounter same issue...

As I mentioned above, memory leak problem solved by changing to another machine and I did not encounter this issue no more. The problem is bijectors take too much ram...