| Portability | POSIX |
|---|---|
| Stability | experimental |
| Maintainer | andrew+cabal@tullo.ch |
| Safe Haskell | None |
DeepLearning.ConvNet
Description
- (>->) :: (Monad m, Shape sh, Shape sh', Shape sh'') => Forward m sh sh' -> Forward m sh' sh'' -> Forward m sh sh''
- type DVol sh = Array D sh Double
- type Forward m sh sh' = Vol sh -> WriterT [Vector Double] m (DVol sh')
- class (Shape sh, Shape sh') => InnerLayer a sh sh' | a -> sh, a -> sh'
- data SoftMaxLayer = SoftMaxLayer
- class TopLayer a
- type Vol sh = Array U sh Double
- flowNetwork :: (Monad m, Shape sh) => sh -> Int -> Int -> Int -> Forward m sh DIM1
- net1 :: (Monad m, InnerLayer a sh DIM1, TopLayer a1) => a -> a1 -> Forward m sh DIM1
- net2 :: (Monad m, InnerLayer a sh sh', InnerLayer a1 sh' DIM1, TopLayer a2) => a -> a1 -> a2 -> Forward m sh DIM1
- newFC :: Shape sh => sh -> Int -> FullyConnectedLayer sh
- withActivations :: Forward m sh sh' -> Vol sh -> m (DVol sh', [Vector Double])
Documentation
(>->) :: (Monad m, Shape sh, Shape sh', Shape sh'') => Forward m sh sh' -> Forward m sh' sh'' -> Forward m sh sh''Source
'(>->)' composes two forward activation functions
type Forward m sh sh' = Vol sh -> WriterT [Vector Double] m (DVol sh')Source
The Forward function represents a single forward pass through a layer.
class (Shape sh, Shape sh') => InnerLayer a sh sh' | a -> sh, a -> sh'Source
InnerLayer represents an inner layer of a neural network that
can accept backpropagation input from higher layers
data SoftMaxLayer Source
SoftMaxLayer computes the softmax activation function.
Constructors
| SoftMaxLayer |
Instances
TopLayer is a top level layer that can initialize a
backpropagation pass.
Instances
flowNetwork :: (Monad m, Shape sh) => sh -> Int -> Int -> Int -> Forward m sh DIM1Source
FlowNetwork builds a network of the form
Input Layer Output Softmax
+--+
| | Inner Layers +--+ +--+
| | | | | |
| | +-+ +-+ +-+ | | | |
| +---+ +---+ +--+ +--+ +---> |
| | +-+ +-+ +-+ | | | |
| | | | | |
| | +--+ +--+
+--+
net1 :: (Monad m, InnerLayer a sh DIM1, TopLayer a1) => a -> a1 -> Forward m sh DIM1Source
net1 constructs a single-layer fully connected perceptron with
softmax output.
net2 :: (Monad m, InnerLayer a sh sh', InnerLayer a1 sh' DIM1, TopLayer a2) => a -> a1 -> a2 -> Forward m sh DIM1Source
net1 constructs a two-layer fully connected MLP with
softmax output.
newFC :: Shape sh => sh -> Int -> FullyConnectedLayer shSource
newFC constructs a new fully connected layer
withActivations :: Forward m sh sh' -> Vol sh -> m (DVol sh', [Vector Double])Source
withActivations computes the output activation, along with the
intermediate activations