MNIST Main compiles but is broken.

tensorflow-mnist/app/Main.hs

@@ -15,7 +15,7 @@
 {-# LANGUAGE FlexibleContexts #-}
 {-# LANGUAGE OverloadedLists #-}
 
-import Control.Monad (zipWithM, when, forM_)
+import Control.Monad (forM_, when)
 import Control.Monad.IO.Class (liftIO)
 import Data.Int (Int32, Int64)
 import Data.List (genericLength)
@@ -30,6 +30,7 @@ import qualified TensorFlow.Ops as TF
 import qualified TensorFlow.Session as TF
 import qualified TensorFlow.Tensor as TF
 import qualified TensorFlow.Types as TF
+import qualified TensorFlow.GenOps.Core as CoreOps
 
 import TensorFlow.Examples.MNIST.InputData
 import TensorFlow.Examples.MNIST.Parse
@@ -64,6 +65,7 @@ data Model = Model {
 
 createModel :: TF.Build Model
 createModel = do
+    let rd = CoreOps.readVariableOp
     -- Use -1 batch size to support variable sized batches.
     let batchSize = -1
     -- Inputs.
@@ -73,13 +75,14 @@ createModel = do
     hiddenWeights <-
         TF.initializedVariable =<< randomParam numPixels [numPixels, numUnits]
     hiddenBiases <- TF.zeroInitializedVariable [numUnits]
-    let hiddenZ = (images `TF.matMul` hiddenWeights) `TF.add` hiddenBiases
+    let hiddenZ = (images `TF.matMul` rd hiddenWeights)
+                  `TF.add` rd hiddenBiases
     let hidden = TF.relu hiddenZ
     -- Logits.
     logitWeights <-
         TF.initializedVariable =<< randomParam numUnits [numUnits, numLabels]
     logitBiases <- TF.zeroInitializedVariable [numLabels]
-    let logits = (hidden `TF.matMul` logitWeights) `TF.add` logitBiases
+    let logits = (hidden `TF.matMul` rd logitWeights) `TF.add` rd logitBiases
     predict <- TF.render $ TF.cast $
                TF.argMax (TF.softmax logits) (TF.scalar (1 :: LabelType))
 
@@ -89,11 +92,11 @@ createModel = do
         loss =
             reduceMean $ fst $ TF.softmaxCrossEntropyWithLogits logits labelVecs
         params = [hiddenWeights, hiddenBiases, logitWeights, logitBiases]
-    grads <- TF.gradients loss params
+    grads <- TF.gradients loss (map rd params)
 
-    let lr = TF.scalar 0.00001
-        applyGrad param grad = TF.assign param $ param `TF.sub` (lr * grad)
-    trainStep <- TF.group =<< zipWithM applyGrad params grads
+    let lr = TF.scalar (-0.00001 :: Float)  -- Negative to make it descend.
+        applyGrad var grad = CoreOps.assignVariableOp var (lr * grad)
+    trainStep <- TF.group (zipWith applyGrad params grads)
 
     let correctPredictions = TF.equal predict labels
     errorRateTensor <- TF.render $ 1 - reduceMean (TF.cast correctPredictions)

tensorflow-mnist/tensorflow-mnist.cabal

@@ -45,6 +45,7 @@ executable Main
                , lens-family
                , proto-lens
                , tensorflow
+               , tensorflow-core-ops
                , tensorflow-mnist
                , tensorflow-mnist-input-data
                , tensorflow-ops

tensorflow-ops/src/TensorFlow/Gradient.hs

@@ -654,6 +654,7 @@ opGrad "ZerosLike" _ _ _ = [Nothing]
 opGrad "Const" _ _ _ = [Nothing, Nothing]
 opGrad "Placeholder" _ _ _ = []
 opGrad "Variable" _ _ _ = []
+opGrad "ReadVariableOp" _ _ _ = [Nothing, Nothing]
 
 opGrad n nodeDef ins grads =
     error $ "no gradient implemented for " ++
@@ -699,6 +700,7 @@ numOutputs o =
         "Transpose" -> 1
         "TruncatedNormal" -> 1
         "Variable" -> 1
+        "ReadVariableOp" -> 2
         "ZerosLike" -> 1
         _ -> error $ "numOuputs not implemented for " ++ show (o ^. op)
 

tensorflow-ops/src/TensorFlow/Ops.hs

@@ -121,8 +121,7 @@ import qualified Proto.Tensorflow.Core.Framework.TensorShape
   as TensorShape
 import TensorFlow.Build
 import TensorFlow.BuildOp
-import TensorFlow.ControlFlow (group)
-import TensorFlow.Output (unNodeName)
+import TensorFlow.Output (ResourceHandle, unNodeName)
 import TensorFlow.Tensor
 import TensorFlow.Types
 
@@ -176,37 +175,30 @@ assign = buildOp $ opDef "Assign"
 -- | Creates a variable initialized to the given value.
 -- Initialization happens next time session runs.
 initializedVariable :: forall a . TensorType a
-                    => Tensor Value a -> Build (Tensor Ref a)
+                    => Tensor Value a -> Build (ResourceHandle a)
 initializedVariable initializer = do
-    v <- variable []  -- The shape is not known initially.
-    (i :: Tensor Ref a) <-
-        buildOp (opDef "Assign"
-                 & opAttr "T" .~ tensorType (undefined :: a)
-                 & opAttr "use_locking" .~ True
-                 & opAttr "validate_shape" .~ False
-                 )
-        v initializer
-    addInitializer =<< group i
+    let v = CoreOps.varHandleOp
+            & resourceHandleAttr "shape" .~ (Shape [])
+    addInitializer (CoreOps.createVariableOp v initializer)
     return v
 
 -- | Creates a zero-initialized variable with the given shape.
 zeroInitializedVariable
-  :: (TensorType a, Num a) =>
-     TensorFlow.Types.Shape -> Build (Tensor TensorFlow.Tensor.Ref a)
+  :: (TensorType a, Num a) => TensorFlow.Types.Shape -> Build (ResourceHandle a)
 zeroInitializedVariable = initializedVariable . zeros
 
 -- TODO: Support heterogeneous list of tensors.
-save :: forall a v . TensorType a
+save :: forall a . TensorType a
         => ByteString     -- ^ File path.
-        -> [Tensor v a]  -- ^ Tensors to save.
+        -> [ResourceHandle a]  -- ^ Tensors to save.
         -> Build ControlNode
 save path xs = do
     let toByteStringTensor = scalar . encodeUtf8 . unNodeName
-    names <- mapM (fmap toByteStringTensor . renderNodeName) xs
+    names <- mapM (fmap toByteStringTensor . renderResourceHandle) xs
     let types = replicate (length xs) (tensorType (undefined :: a))
     let saveOp = buildOp $ opDef "Save"
                          & opAttr "T" .~ types
-    saveOp (scalar path) (CoreOps.pack names) xs
+    saveOp (scalar path) (CoreOps.pack names) (map CoreOps.readVariableOp xs)
 
 -- | Restore a tensor's value from a checkpoint file.
 --
@@ -215,21 +207,21 @@ save path xs = do
 restoreFromName :: forall a . TensorType a
                 => ByteString    -- ^ File path.
                 -> ByteString    -- ^ Tensor name override.
-                -> Tensor Ref a  -- ^ Tensor to restore.
-                -> Build ControlNode
-restoreFromName path name x = do
+                -> Build (Tensor Value a)
+restoreFromName path name = do
     let restoreOp = buildOp $ opDef "Restore"
                             & opAttr "dt" .~ tensorType (undefined :: a)
-    group =<< assign x (restoreOp (scalar path) (scalar name) :: Tensor Value a)
+    restoreOp (scalar path) (scalar name)
+
 
 -- | Restore a tensor's value from a checkpoint file.
 restore :: forall a . TensorType a
         => ByteString    -- ^ File path.
-        -> Tensor Ref a  -- ^ Tensor to restore.
+        -> ResourceHandle a
         -> Build ControlNode
 restore path x = do
-    name <- encodeUtf8 . unNodeName <$> renderNodeName x
-    restoreFromName path name x
+    name <- encodeUtf8 . unNodeName <$> renderResourceHandle x
+    CoreOps.assignVariableOp x <$> restoreFromName path name
 
 -- | Create a constant tensor.
 --
@@ -253,12 +245,13 @@ constant (Shape shape') values
     typedNode :: TensorProto
     typedNode = def
                 & dtype .~ nodeType
+                -- Use shapeToProto from Types.hs
                 & tensorShape.TensorShape.dim .~
                       [def & TensorShape.size .~ x | x <- shape']
                 & tensorVal .~ values
 
 -- | Reshape a N-D tensor down to a scalar.
--- 
+--
 -- See `TensorFlow.GenOps.Core.reshape`.
 scalarize :: (TensorType a) => Tensor v a -> Tensor Value a
 scalarize t = CoreOps.reshape t (vector scalarShape)

tensorflow-ops/tensorflow-ops.cabal

@@ -41,6 +41,7 @@ Test-Suite BuildTest
                , lens-family
                , google-shim
                , tensorflow
+               , tensorflow-core-ops
                , tensorflow-ops
                , tensorflow-proto
                , test-framework

tensorflow-ops/tests/BuildTest.hs

@@ -64,6 +64,7 @@ import Test.Framework.Providers.HUnit (testCase)
 import Test.HUnit ((@=?))
 import Google.Test (googleTest)
 import qualified Data.Vector as V
+import qualified TensorFlow.GenOps.Core as CoreOps
 
 -- | Test named behavior.
 testNamed :: Test
@@ -96,13 +97,12 @@ testPureRender = testCase "testPureRender" $ runSession $ do
 testInitializedVariable :: Test
 testInitializedVariable =
     testCase "testInitializedVariable" $ runSession $ do
-        (formula, reset) <- build $ do
+        (v, formula) <- build $ do
             v <- initializedVariable 42
-            r <- assign v 24
-            return (1 `add` v, r)
+            return (v, 1 `add` CoreOps.readVariableOp v)
         result <- run formula
         liftIO $ 43 @=? (unScalar result :: Float)
-        run_ reset  -- Updates v to a different value
+        run_ (CoreOps.assignVariableOp v 24)
         rerunResult <- run formula
         liftIO $ 25 @=? (unScalar rerunResult :: Float)
 
@@ -110,7 +110,7 @@ testInitializedVariableShape :: Test
 testInitializedVariableShape =
     testCase "testInitializedVariableShape" $ runSession $ do
         vector <- build $ initializedVariable (constant [1] [42 :: Float])
-        result <- run vector
+        result <- run (CoreOps.readVariableOp vector)
         liftIO $ [42] @=? (result :: V.Vector Float)
 
 -- | Test nameScoped behavior.

tensorflow-ops/tests/EmbeddingOpsTest.hs

@@ -112,8 +112,9 @@ testEmbeddingLookupGradients = testCase "testEmbeddingLookupGradients" $ do
             let ids           = TF.constant (TF.Shape [1, 2]) idValues
 
             x <- TF.placeholder (TF.Shape [2])
-            embedding <- TF.initializedVariable
-                            =<< TF.render (TF.constant embShape embeddingInit)
+            embedding <- CoreOps.readVariableOp <$>
+                         (TF.initializedVariable
+                          =<< TF.render (TF.constant embShape embeddingInit))
 
             op <- embeddingLookup [embedding] ids
             let twoNorm = CoreOps.square $ TF.abs (op - x)

tensorflow-ops/tests/OpsTest.hs

@@ -28,10 +28,11 @@ import qualified Data.ByteString.Char8 as B8
 import qualified Data.Vector as V
 import qualified TensorFlow.Build as TF
 import qualified TensorFlow.ControlFlow as TF
+import qualified TensorFlow.GenOps.Core as CoreOps
 import qualified TensorFlow.Nodes as TF
 import qualified TensorFlow.Ops as TF
+import qualified TensorFlow.Output as TF
 import qualified TensorFlow.Session as TF
-import qualified TensorFlow.Tensor as TF
 import qualified TensorFlow.Types as TF
 
 -- | Test that one can easily determine number of elements in the tensor.
@@ -54,17 +55,16 @@ testSaveRestore :: Test
 testSaveRestore = testCase "testSaveRestore" $
     withSystemTempDirectory "" $ \dirPath -> do
         let path = B8.pack $ dirPath ++ "/checkpoint"
-            var :: TF.Build (TF.Tensor TF.Ref Float)
-            var = TF.render =<<
-                  TF.named "a" <$> TF.zeroInitializedVariable (TF.Shape [])
+            var :: TF.Build (TF.ResourceHandle Float)
+            var = TF.zeroInitializedVariable (TF.Shape [])
         TF.runSession $ do
             v <- TF.build var
-            TF.buildAnd TF.run_ $ TF.assign v 134
+            TF.buildAnd TF.run_ $ TF.group $ CoreOps.assignVariableOp v 134
             TF.buildAnd TF.run_ $ TF.save path [v]
         result <- TF.runSession $ do
             v <- TF.build var
             TF.buildAnd TF.run_ $ TF.restore path v
-            TF.run v
+            TF.run (CoreOps.readVariableOp v)
         liftIO $ TF.Scalar 134 @=? result
 
 

tensorflow/src/TensorFlow/Build.hs

@@ -34,6 +34,7 @@ module TensorFlow.Build
     , GraphState
     , render
     , renderNodeName
+    , renderResourceHandle
     , renderedNodeDefs
     , BuildT
     , Build
@@ -361,6 +362,9 @@ render = tensorOutput $ outputOp $ fmap Rendered . resolveOp
 renderNodeName :: Tensor v a -> Build NodeName
 renderNodeName t = getOrAddOp (t ^. tensorOutput . outputOp)
 
+renderResourceHandle :: ResourceHandle a -> Build NodeName
+renderResourceHandle (ResourceHandle r) = getOrAddOp (r ^. outputOp)
+
 -- | Records the given summary action in Build for retrieval with
 -- 'collectAllSummaries'. The summary op is required to produce a
 -- Summary protocol buffer in string form. For safety, use the

tensorflow/src/TensorFlow/Output.hs

@@ -38,6 +38,7 @@ module TensorFlow.Output
     , outputOp
     , PendingNodeName(..)
     , ResourceHandle(..)
+    , resourceHandleOutput
     )  where
 
 import qualified Data.Map.Strict as Map
@@ -160,4 +161,7 @@ instance IsString Output where
 -- | Opaque handle to a mutable resource in the graph.  Typical such
 -- resources are variables. The type parameter corresponds to the
 -- dtype of the tensor held in the variable.
-newtype ResourceHandle a = ResourceHandle Output
+newtype ResourceHandle a = ResourceHandle { unResourceHandle :: Output }
+
+resourceHandleOutput :: Lens' (ResourceHandle a) Output
+resourceHandleOutput = lens unResourceHandle (\_ x -> ResourceHandle x)

tensorflow/src/TensorFlow/Tensor.hs

@@ -24,7 +24,14 @@ import qualified Data.Text as Text
 import Lens.Family2 (Lens', Traversal')
 import Lens.Family2.Unchecked (lens)
 
-import TensorFlow.Output (Output, outputOp, opUnrendered, opAttr)
+import TensorFlow.Output
+    ( Output
+    , ResourceHandle
+    , outputOp
+    , opUnrendered
+    , opAttr
+    , resourceHandleOutput
+    )
 import TensorFlow.Types (TensorData(..), Attribute)
 import qualified TensorFlow.Internal.FFI as FFI
 
@@ -61,6 +68,11 @@ tensorOutput = lens (\(Tensor _ o) -> o) (\(Tensor v _) o -> Tensor v o)
 tensorAttr :: Attribute attr => Text.Text -> Traversal' (Tensor v a) attr
 tensorAttr x = tensorOutput . outputOp . opUnrendered . opAttr x
 
+resourceHandleAttr :: Attribute attr
+                      => Text.Text -> Traversal' (ResourceHandle a) attr
+resourceHandleAttr attr =
+    resourceHandleOutput . outputOp . opUnrendered . opAttr attr
+
 -- | Cast a 'Tensor *' into a 'Tensor Value'. Common usage is to cast a
 -- Ref into Value. This behaves like a no-op.
 value :: Tensor v a -> Tensor Value a