Multi-dimensional Arrays

This is basically multi-dimensional data. So include large $n\times m\times o\times p$ matrices. Could be accelerated by GPU. Also could be used for neural networks.
Perhaps a multi-dimensional array type may be defined as: MultiArray(Real,[3,4,5]) PRINT SIMP SIMP ONCE FLOAT EQ GRAPH DEFEVAL EXPAND NATS And we would set data by: MultiArray.fromList(Real,[2,2], [Type.id(Real)(1),2,3,4] ) PRINT SIMP SIMP ONCE FLOAT EQ GRAPH DEFEVAL EXPAND NATS We would be able to display these tensors in the canvas a= MultiArray.fromFuncL(Real, [3,50,50], FUN(List(Real), p=> Real.Step( Real.sub(Real.plus(Real.power(List.get(Real,1,p),2 ), Real.power(List.get(Real,2,p) ,2)), 2500) ) )) ; plotBitmap(3,50,50,a.float()) PRINT SIMP SIMP ONCE FLOAT EQ GRAPH DEFEVAL EXPAND NATS
For fast array manipulation on GPU we want to act directly on the arrays. e.g. $Array.Sin(F)$

Loading Bitmaps

Neural Networks

A simple neural network which involes a fully connected layer and a tanh operation: $$\sqrt{1}$$