added default handlers documentation

docs/default_handlers.md

@@ -0,0 +1,36 @@
+# TorchServe default inference handlers
+
+TorchServe provides following inference handlers out of box :
+
+## image_classifier
+
+ * Description : Handles image classification models trained on imagenet dataset.
+ * Input : RGB image
+ * Output : Top 5 predictions and their respective probability of the image
+
+For more details refer [examples](../examples/image_classifier)
+
+## image_segmenter
+
+ * Description : Handles image segmentation models trained on imagenet dataset.
+ * Input : RGB image
+ * Output : Output shape as [CL H W], CL - number of classes, H - height and W - width.
+
+For more details refer [examples](../examples/image_segmenter)
+
+## object_detector
+
+ * Description : Handles object detection models.
+ * Input : RGB image
+ * Output : List of detected classes and bounding boxes respectively 
+
+For more details refer [examples](../examples/object_detector) 
+
+## text_classifier
+
+ * Description : Handles models trained on imagenet dataset.
+ * Input : text file
+ * Output : Class of input text
+
+For more details refer [examples](../examples/text_classification)
+