backend: Support graph computation that does not return an output

There are two cases where we may not have an output after computing:
 - Prompt processing where the length of the input exceeds the batch
   size
 - Internal memory management operations such as cache defrag and shift

model/model.go

@@ -275,5 +275,8 @@ func Forward(m Model, optsFuncs ...OptionsFunc) (ml.Tensor, error) {
 	}
 	defer ctx.Close()
 
-	return ctx.Compute(t), nil
+	ctx.Forward(t)
+	ctx.Compute(t)
+
+	return t, nil
 }