Anssi

@Waterkin Can you share more details? - Which code did you try to run? - What kind of system are you running (OS, library versions, etc)

Ah gotcha! Sadly I am not a mac person so can not comment on the bugs too much (sounds like you found something that needs to be better documented, so...

ViZDoom does not currently support that, and as far I have understood ZDoom's innards this would be rather difficult to setup. However I wonder if demos/replays could be used for...

@aklein1995 Unfortunately it has not been implemented yet, but we would be happy to accept PRs that add such functionality if anybody has time to implement it :)

Learning examples use Q-learning with replay buffer, which use a lot of memory, especially with images. You can try setting the `replay_memory_size` parameter to something lower, but this can make...

> In that case, should memory be full since the early epoch? Not necessarily. Depending on how the buffer is created, it might slowly fill up as training goes. This...

I ran the code for 20min on a Ubuntu 20.04 machine with Python 3.7, ViZDoom 1.1.8 and PyTorch 1.8.1, and the memory usage maxed out at 9GB for me (for...

> I ran it on 64GB ram. So, there may be something wrong with the system or software, not my code? The sudden spike seems to indicate so. However now...

Try commenting out [this line](https://github.com/mwydmuch/ViZDoom/blob/master/examples/python/basic.py#L87). It seems the error stems from audio library, which is not very essential to the use of ViZDoom (only if you want humans to play...

@mwydmuch I am currently working on digging into this (see some relevant discussions I found with googling below). As far I can see the issue does not directly come from...