Generate tflite model format as well along with .h5

backend/core/tasks.py

@@ -222,6 +222,17 @@ def train_model(
             # Save the model in HDF5 format
             model.save(os.path.join(output_path, "checkpoint.h5"))
 
+            logger.info(model.inputs)
+            logger.info(model.outputs)
+
+            # Convert the model to tflite for android/ios.
+            converter = tf.lite.TFLiteConverter.from_keras_model(model)
+            tflite_model = converter.convert()
+
+            # Save the model.
+            with open(os.path.join(output_path, "checkpoint.tflite"), 'wb') as f:
+                f.write(tflite_model)
+
             # now remove the ramp-data all our outputs are copied to our training workspace
             shutil.rmtree(base_path)
             training_instance.accuracy = float(final_accuracy)