Gabriel Previato
Gabriel Previato
I'm having the same issue with Julia v1.5.2 and ONNX.jl v0.1.1
Same problem here. Just found out that the problem is that the PyPI version of the AirSim is 1.7.0. In the 1.8.1 they added a new parameter for both simSetObjectMaterial...
@Dongzhou-1996 just update the package, it should work now.
I added a PR that should fix this, https://github.com/FluxML/NNlib.jl/pull/468.
Hum, I think I've read somewhere something like "half precision support", but indeed, BLAS only supports 32/64. But still, if you mix Float32 and Float16 (it was what happened to...
I don't think we should be dissuading people from mixing types or doing weird combinations, but I think the warning is something valid and useful once you can actually do...
If I got it right, the package was fixed to be 1.0 compatible, but it was not released yet, since the way packages are released now has been changed, and...
Yes, this would work for the weights in all labels, but if you want it for the positive labels it wouldn't work, right? Something like this for LogitBCE: ```julia function...
Nice one! I think it would be good to have it there. From my experience, coming from other Python frameworks they have it built-in, so I think showing this more...
One solution that I thought is breaking bigger arrays in smaller chunks. Something like this: ```julia new_mean(A::AbstractArray{ 10000 chunks = [collect(1:10000:length(A)); length(A)] A_views = [view(A, chunks[i]:chunks[i+1]) for i in 1:length(chunks)-1]...