fixed #88 - example/char-rnn memory leak

CHANGELOG.md

@@ -10,6 +10,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Added `nn.MSELoss()`
 - reworked `ts.Format()`
 - Added conv2d benchmark
+- Fixed #88 memory leak at `example/char-rnn`
 
 ## [Nofix]
 - ctype `long` caused compiling error in MacOS as noted on [#44]. Not working on linux box.

example/char-rnn/README.md

@@ -11,7 +11,7 @@ At the end of each training epoch, some sample text is generated and printed.
 Any text file can be used as an input, as long as it's large enough for training.
 A typical example would be the
 [tiny Shakespeare dataset](https://raw.githubusercontent.com/karpathy/char-rnn/master/data/tinyshakespeare/input.txt).
-The training text file should be stored in `data/input.txt`.
+The training text file should be stored in `data/char-rnn/input.txt`.
 
 To run the example:
 

example/char-rnn/main.go

@@ -14,12 +14,11 @@ const (
 	HiddenSize   int64   = 256
 	SeqLen       int64   = 180
 	BatchSize    int64   = 256
-	Epochs       int     = 100
+	Epochs       int     = 3
 	SamplingLen  int64   = 1024
 )
 
 func sample(data *ts.TextData, lstm *nn.LSTM, linear *nn.Linear, device gotch.Device) string {
-
 	labels := data.Labels()
 	inState := lstm.ZeroState(1)
 	lastLabel := int64(0)
@@ -59,13 +58,12 @@ func sample(data *ts.TextData, lstm *nn.LSTM, linear *nn.Linear, device gotch.De
 }
 
 func main() {
-	cuda := gotch.NewCuda()
+	device := gotch.CudaIfAvailable()
-	device := cuda.CudaIfAvailable()
 
 	vs := nn.NewVarStore(device)
 	data, err := ts.NewTextData("../../data/char-rnn/input.txt")
 	if err != nil {
-		log.Fatal(err)
+		panic(err)
 	}
 
 	labels := data.Labels()
@@ -94,41 +92,36 @@ func main() {
 			}
 
 			batchNarrow := batchTs.MustNarrow(1, 0, SeqLen, false)
-			xsOnehotTmp := batchNarrow.Onehot(labels)
+			xsOnehot := batchNarrow.Onehot(labels).MustTo(device, true) // [256, 180, 65]
-			xsOnehot := xsOnehotTmp.MustTo(device, true) // shape: [256, 180, 65]
+			batchNarrow.MustDrop()
-			ysTmp1 := batchTs.MustNarrow(1, 1, SeqLen, true)
+
-			ysTmp2 := ysTmp1.MustTotype(gotch.Int64, true)
+			ys := batchTs.MustNarrow(1, 1, SeqLen, true).MustTotype(gotch.Int64, true).MustTo(device, true).MustView([]int64{BatchSize * SeqLen}, true)
-			ysTmp3 := ysTmp2.MustTo(device, true)
-			ys := ysTmp3.MustView([]int64{BatchSize * SeqLen}, true)
 
 			lstmOut, outState := lstm.Seq(xsOnehot)
 			// NOTE. Although outState will not be used. There a hidden memory usage
 			// on C land memory that is needed to free up. Don't use `_`
 			outState.(*nn.LSTMState).Tensor1.MustDrop()
 			outState.(*nn.LSTMState).Tensor2.MustDrop()
+			xsOnehot.MustDrop()
 
 			logits := linear.Forward(lstmOut)
+			lstmOut.MustDrop()
 			lossView := logits.MustView([]int64{BatchSize * SeqLen, labels}, true)
 
 			loss := lossView.CrossEntropyForLogits(ys)
+			ys.MustDrop()
+			lossView.MustDrop()
 
 			opt.BackwardStepClip(loss, 0.5)
 			sumLoss += loss.Float64Values()[0]
 			cntLoss += 1.0
-
-			// batchTs.MustDrop()
-			// batchNarrow.MustDrop()
-			// xsOnehotTmp.MustDrop()
-			xsOnehot.MustDrop()
-			ys.MustDrop()
-			lstmOut.MustDrop()
 			loss.MustDrop()
 
 			batchCount++
 			if batchCount%500 == 0 {
 				fmt.Printf("Epoch %v - Batch %v \n", epoch, batchCount)
 			}
-		}
+		} // infinite for-loop
 
 		sampleStr := sample(data, lstm, linear, device)
 		fmt.Printf("Epoch %v - Loss: %v \n", epoch, sumLoss/cntLoss)
@@ -137,5 +130,4 @@ func main() {
 		dataIter.Data.MustDrop()
 		dataIter.Indexes.MustDrop()
 	}
-
 }