# DimensionMismatch: matrix A has dimensions (100,10), matrix B has dimensions (1,7)
**URL:** https://discourse.julialang.org/t/dimensionmismatch-matrix-a-has-dimensions-100-10-matrix-b-has-dimensions-1-7/106528
**Category:** New to Julia
**Tags:** question, neural-network
**Created:** [November 21, 2023, 1:51pm UTC](https://discourse.julialang.org/t/dimensionmismatch-matrix-a-has-dimensions-100-10-matrix-b-has-dimensions-1-7/106528 "2023-11-21T13:51:02Z")
**Posts on this page:** 6
**Page:** 1
### Author:  [@SimonHubert](https://discourse.julialang.org/u/SimonHubert)
#### Post date: [November 21, 2023, 1:51pm UTC](https://discourse.julialang.org/t/dimensionmismatch-matrix-a-has-dimensions-100-10-matrix-b-has-dimensions-1-7/106528/1 "2023-11-21T13:51:02Z")
Hello I have this problem with my neural network. I want to classify a 7 segment led digits.
Problem is i get this error:
```julia
DimensionMismatch: matrix A has dimensions (100,10), matrix B has dimensions (1,7)
1. **_generic_matmatmul!** (::Matrix{Float32}, ::Char, ::Char, ::Matrix{Float32}, ::Vector{Int64}, ::LinearAlgebra.MulAddMul{true, true, Bool, Bool})@*matmul.jl:856*
2. **generic_matmatmul!** @*matmul.jl:847* [inlined]
3. **mul!** @*matmul.jl:407* [inlined]
4. **mul!** @*matmul.jl:276* [inlined]
5. ***** @*matmul.jl:141* [inlined]
6. **rrule** @*arraymath.jl:40* [inlined]
7. **rrule** @*rules.jl:134* [inlined]
8. **chain_rrule** @*chainrules.jl:223* [inlined]
9. **macro expansion** @*interface2.jl:101* [inlined]
10. **_pullback** @*interface2.jl:101* [inlined]
11. **_pullback** @*[Other: 10](http://localhost:1234/edit?id=f53566dc-8710-11ee-3282-15008b18595c#)* [inlined]
12. **_pullback** (::Zygote.Context{true}, ::Main.var"workspace#4".Layer, ::LinearAlgebra.Transpose{Int64, Vector{Int64}})@*interface2.jl:0*
13. **_apply** @*boot.jl:838* [inlined]
14. **adjoint** @*lib.jl:203* [inlined]
15. **_pullback** @*adjoint.jl:66* [inlined]
16. **_pullback** @*operators.jl:1035* [inlined]
17. **_pullback** @*operators.jl:1034* [inlined]
18. **_pullback** @*operators.jl:1031* [inlined]
19. **_pullback** (::Zygote.Context{true}, ::Base.var"##_#97", ::Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}, ::ComposedFunction{Main.var"workspace#4".Layer, Main.var"workspace#4".Layer}, ::LinearAlgebra.Transpose{Int64, Vector{Int64}})@*interface2.jl:0*
20. **_apply** (::Function, ::Vararg{Any})@*boot.jl:838*
21. **adjoint** @*lib.jl:203* [inlined]
22. **_pullback** @*adjoint.jl:66* [inlined]
23. **_pullback** @*operators.jl:1031* [inlined]
24. **_pullback** (::Zygote.Context{true}, ::ComposedFunction{Main.var"workspace#4".Layer, Main.var"workspace#4".Layer}, ::LinearAlgebra.Transpose{Int64, Vector{Int64}})@*interface2.jl:0*
25. **_pullback** @*[Other: 8](http://localhost:1234/edit?id=f53566dc-8710-11ee-3282-15008b18595c#)* [inlined]
26. **_pullback** (::Zygote.Context{true}, ::Main.var"workspace#124".Network, ::LinearAlgebra.Transpose{Int64, Vector{Int64}})@*interface2.jl:0*
27. **_pullback** @*[Local: 18](http://localhost:1234/edit?id=f53566dc-8710-11ee-3282-15008b18595c#)* [inlined]
28. **_pullback** (::Zygote.Context{true}, ::Main.var"workspace#2623".var"#1#2"{Int64})@*interface2.jl:0*
29. **pullback** (::Function, ::Zygote.Params{Zygote.Buffer{Any, Vector{Any}}})@*interface.jl:414*
30. **gradient** (::Function, ::Zygote.Params{Zygote.Buffer{Any, Vector{Any}}})@*interface.jl:96*
31. **top-level scope** @*[Local: 17](http://localhost:1234/edit?id=f53566dc-8710-11ee-3282-15008b18595c#)*
```
Here is my code to this problem:
```julia
begin
struct Layer
W::Matrix{Float32} # weight matrix - Float32 for faster gradients
b::Vector{Float32} # bias vector
activation::Function
Layer(in::Int64,out::Int64,activation::Function=identityFunction) =
new(randn(out,in),randn(out),activation) # constructor
end
(m::Layer)(x) = m.activation.(m.W*x .+ m.b) # feed-forward pass
end
begin
ReLu(x) = max(0,x)
identityFunction(x) = x
end;
begin
struct Network
layers::Vector{Layer}
Network(layers::Vararg{Layer}) = new(vcat(layers...))
# constructor - allow arbitrarily many layers
end
(n::Network)(x) = reduce((left,right)->right∘left, n.layers)(x)
# perform layer-wise operations over arbitrarily many layers
end
begin
inputs = [
1 1 1 1 1 1 0;
0 1 1 0 0 0 0;
1 1 0 1 1 0 1;
1 1 1 1 0 0 1;
0 1 1 0 0 1 1;
1 0 1 1 0 1 1;
1 0 1 1 1 1 1;
1 1 1 0 0 0 0;
1 1 1 1 0 0 1;
1 1 1 1 0 1 1
]
# create training data
targetOutput = [
1 0 0 0 0 0 0 0 0 0;
0 1 0 0 0 0 0 0 0 0;
0 0 1 0 0 0 0 0 0 0;
0 0 0 1 0 0 0 0 0 0;
0 0 0 0 1 0 0 0 0 0;
0 0 0 0 0 1 0 0 0 0;
0 0 0 0 0 0 1 0 0 0;
0 0 0 0 0 0 0 1 0 0;
0 0 0 0 0 0 0 0 1 0;
0 0 0 0 0 0 0 0 0 1
]
mse(x,y) = sum((x .- y).^2)/length(x) # MSE will be our loss function
using Random
Random.seed!(54321) # for reproducibility
twoLayerNeuralNet = Network(Layer(10,100,ReLu), Layer(100,10)) # instantiate a two-layer network
end
begin
# Packages for automatic differentiation and neural networks
# (i.e. Tensorflow for Julia)
Flux.@functor Layer # set the Layer-struct as being differentiable
Flux.@functor Network # set the Network-struct as being differentiable
parameters = Flux.params(twoLayerNeuralNet)
# obtain the parameters of the layers (recurses through network)
optimizer = ADAM(0.01) # from Flux-library
netOutput = [] # store output for plotting
lossCurve = [] # store loss for plotting
for i in 1:500
for j in shuffle(0:9)
# Calculate the gradients for the network parameters
gradients = Zygote.gradient(
() -> mse(
twoLayerNeuralNet(transpose(inputs[j,:]))[:],
targetOutput[j,:]),
parameters
)
# Update the parameters using the gradients and optimiser settings.
Flux.Optimise.update!(optimizer, parameters, gradients)
# Log the performance for later plotting
actualOutput = twoLayerNeuralNet(transpose(inputs[j,:]))[:]
push!(netOutput, actualOutput)
push!(lossCurve,
mse(
actualOutput,
targetOutput[j,:]))
end
end
end
```
I dont know how to resolve the problem. Help would be appreciated
---
### Author:  [@Dan](https://discourse.julialang.org/u/Dan)
#### Post date: [November 21, 2023, 2:18pm UTC](https://discourse.julialang.org/t/dimensionmismatch-matrix-a-has-dimensions-100-10-matrix-b-has-dimensions-1-7/106528/2 "2023-11-21T14:18:21Z")
The `input`s are vectors with **7** features, yet the network is defined from **10** inputs. Maybe changing:
```julia
Network(Layer(10,100,ReLu), Layer(100,10))
```
to
```julia
Network(Layer(7,100,ReLu), Layer(100,10))
```
would help.
---
### Author:  [@SimonHubert](https://discourse.julialang.org/u/SimonHubert)
#### Post date: [November 21, 2023, 2:28pm UTC](https://discourse.julialang.org/t/dimensionmismatch-matrix-a-has-dimensions-100-10-matrix-b-has-dimensions-1-7/106528/3 "2023-11-21T14:28:14Z")
@Dan Now the error is [DimensionMismatch: matrix A has dimensions (100,7), matrix B has dimensions (1,7)]. I used 10 because i wanted to randomize the input nodes. So that it starts in every iteration with a different randomized node.
---
### Author:  [@Dan](https://discourse.julialang.org/u/Dan)
#### Post date: [November 21, 2023, 2:33pm UTC](https://discourse.julialang.org/t/dimensionmismatch-matrix-a-has-dimensions-100-10-matrix-b-has-dimensions-1-7/106528/4 "2023-11-21T14:33:12Z")
Perhaps also replace:
```julia
transpose(inputs[j,:])
```
with
```julia
vec(inputs[j,:])
```
On a “meta” note, this same question appears also on StackOverflow, twice under two different names, isn’t asking once enough? Or once on each platform.
---
### Author:  [@SimonHubert](https://discourse.julialang.org/u/SimonHubert)
#### Post date: [November 21, 2023, 2:35pm UTC](https://discourse.julialang.org/t/dimensionmismatch-matrix-a-has-dimensions-100-10-matrix-b-has-dimensions-1-7/106528/5 "2023-11-21T14:35:00Z")
Yeah,thats true,i asked on Stack Overflow once. Idk about the other post. Still thank you for your help!
---
### Author:  [@Francis](https://discourse.julialang.org/u/Francis)
#### Post date: [November 25, 2023, 11:17pm UTC](https://discourse.julialang.org/t/dimensionmismatch-matrix-a-has-dimensions-100-10-matrix-b-has-dimensions-1-7/106528/6 "2023-11-25T23:17:18Z")
@SimonHubert Did you get help? I am also having a similar problem