Mocha.jl
Mocha.jl copied to clipboard
Deep Learning framework for Julia
Hi, I was wondering if anyone could shed some light on the following questions regarding the [wasserstein layer](https://github.com/pluskid/Mocha.jl/blob/5e15b882d7dd615b0c5159bb6fde2cc040b2d8ee/src/layers/wasserstein-loss.jl#L50). * How is `M` computed, where can I find this info? *...
When run Pkg.test("Mocha"), I receive3d: ERROR: LoadError: UndefVarError: srand not defined please advise
Based on #253 from @ianshmean - [x] Pass unit tests on CPU - [ ] Make sure examples run on CPU - [ ] Pass unit tests on GPU -...
ERROR: LoadError: LoadError: UndefVarError: is_apple not defined As of Julia v7.0, Base.is_apple() seems to be replaced with Base.Sys.isapple()
Ref: https://discourse.julialang.org/t/package-compatibility-caps/15301
[ Info: Precompiling Mocha [f17d6557-5fdd-57bf-a30c-27e301b4ff87] ERROR: LoadError: InitError: ArgumentError: Unknown time zone "Asia/Shanghai" Stacktrace: [1] (::getfield(TimeZones, Symbol("##1#3")){SubString{String}})() at /Users/marxia/.julia/packages/TimeZones/wytr8/src/TimeZones.jl:70 [2] get!(::getfield(TimeZones, Symbol("##1#3")){SubString{String}}, ::Dict{AbstractString,Dates.TimeZone}, ::SubString{String}) at ./dict.jl:453 [3] Type at /Users/marxia/.julia/packages/TimeZones/wytr8/src/TimeZones.jl:64...
Hello! It could be make pooling without padding? So I would like in pool2: (45\*45\*32\*100) -> (22\*22\*32\*100) 
When building Mocha, an error shows: ``` Running "g++ -fPIC -Wall -03 -shared -fopenmp -o libmochaext.so im2col.cpp pooling.cpp" im2col.cpp:1:0: warning: -fPIC ignored for target #include ^ g++: error: libgomp.spec: No...
Is there any way to make a prediction on a network other than defining a new data layer, creating a network, loading the network from a snapshot, than forwarding it?...
I'm constantly getting "out of memory" error while doing a grid search. It seems like Mocha maps tons of memory for GPU backend. If I'm using a CPU backend everything...