Added tensorflow linear regression toy example

Author: mandar2812

scripts/tf_linear_regression.sc

@@ -0,0 +1,46 @@
+{
+  import org.platanios.tensorflow.api._
+
+  import scala.collection.mutable.ArrayBuffer
+  import scala.util.Random
+
+  val random = new Random()
+
+  val weight = random.nextFloat()
+
+  def batch(batchSize: Int): (Tensor, Tensor) = {
+    val inputs = ArrayBuffer.empty[Float]
+    val outputs = ArrayBuffer.empty[Float]
+    var i = 0
+    while (i < batchSize) {
+      val input = random.nextFloat()
+      inputs += input
+      outputs += weight * input
+      i += 1
+    }
+    (Tensor(inputs).reshape(Shape(-1, 1)), Tensor(outputs).reshape(Shape(-1, 1)))
+  }
+
+  print("Building linear regression model.")
+  val inputs = tf.placeholder(FLOAT32, Shape(-1, 1))
+  val outputs = tf.placeholder(FLOAT32, Shape(-1, 1))
+  val weights = tf.variable("weights", FLOAT32, Shape(1, 1), tf.ZerosInitializer)
+  val predictions = tf.matmul(inputs, weights)
+  val loss = tf.sum(tf.square(predictions - outputs))
+  val trainOp = tf.train.AdaGrad(1.0).minimize(loss)
+
+  println("Training the linear regression model.")
+  val session = Session()
+  session.run(targets = tf.globalVariablesInitializer())
+  for (i <- 0 to 50) {
+    val trainBatch = batch(10000)
+    val feeds = Map(inputs -> trainBatch._1, outputs -> trainBatch._2)
+    val trainLoss = session.run(feeds = feeds, fetches = loss, targets = trainOp)
+    if (i % 1 == 0)
+      println(s"Train loss at iteration $i = ${trainLoss.scalar} " +
+        s"(weight = ${session.run(fetches = weights.value).scalar})")
+  }
+
+  println(s"Trained weight value: ${session.run(fetches = weights.value).scalar}")
+  println(s"True weight value: $weight")
+}