tflite readme updated

anshkumar · May 28, 2022 · fec8e82 · fec8e82
1 parent 97ebcfa
commit fec8e82
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 1 deletion.
diff --git a/README.md b/README.md
@@ -188,6 +188,9 @@ python train.py -tfrecord_train_dir 'path of TFRecord training files'
 # Inference 
 Inside `saved_models` there will be saved graphs according to the score of their validation. For an example `saved_model_0.19968511` is saved_model when the validation loss was 0.19968511. To run inference on using this saved_model see `infer.py`.
 
+# TFLite support
+Please check `to_tflite.py` for converting `saved_mode.pb` to `.tflite` file. 
+
 # Reference
 * https://github.com/feiyuhuahuo/Yolact_minimal
 * https://github.com/dbolya/yolact

diff --git a/to_tflite.py b/to_tflite.py
@@ -44,7 +44,7 @@ def representative_dataset():
             tf.lite.OpsSet.SELECT_TF_OPS # enable TensorFlow ops.
             ]
 converter.inference_input_type = tf.uint8  # or tf.uint8
-converter.inference_output_type = tf.uint8  # or tf.uint8
+# converter.inference_output_type = tf.uint8  # or tf.uint8
 tflite_quant_model = converter.convert()
 
 open("yolact.tflite", "wb").write(tflite_quant_model)