onnx2tf 1.26.1__tar.gz → 1.26.3__tar.gz

{onnx2tf-1.26.1/onnx2tf.egg-info → onnx2tf-1.26.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: onnx2tf
-Version: 1.26.1
+Version: 1.26.3
 Summary: Self-Created Tools to convert ONNX files (NCHW) to TensorFlow/TFLite/Keras format (NHWC). The purpose of this tool is to solve the massive Transpose extrapolation problem in onnx-tensorflow (onnx-tf).
 Home-page: https://github.com/PINTO0309/onnx2tf
 Author: Katsuya Hyodo
@@ -314,7 +314,7 @@ Video speed is adjusted approximately 50 times slower than actual speed.
   docker run --rm -it \
   -v `pwd`:/workdir \
   -w /workdir \
-  ghcr.io/pinto0309/onnx2tf:1.26.1
+  ghcr.io/pinto0309/onnx2tf:1.26.3
 
   or
 
@@ -322,7 +322,7 @@ Video speed is adjusted approximately 50 times slower than actual speed.
   docker run --rm -it \
   -v `pwd`:/workdir \
   -w /workdir \
-  docker.io/pinto0309/onnx2tf:1.26.1
+  docker.io/pinto0309/onnx2tf:1.26.3
 
   or
 
@@ -1463,9 +1463,13 @@ For example, take a model with multiple inputs and multiple outputs as shown in
 When converting to TensorFlow.js, process as follows.
 
 ```bash
-pip install tensorflowjs
+pip install -U --no-deps \
+tensorflowjs \
+tensorflow_decision_forests \
+ydf \
+tensorflow_hub
 
-onnx2tf -i mobilenetv2-12.onnx -ois input:1,3,224,224 -osd
+onnx2tf -i mobilenetv2-12.onnx -ois input:1,3,224,224 -osd -dgc
 
 tensorflowjs_converter \
 --input_format tf_saved_model \

{onnx2tf-1.26.1 → onnx2tf-1.26.3}/README.md

@@ -299,7 +299,7 @@ Video speed is adjusted approximately 50 times slower than actual speed.
   docker run --rm -it \
   -v `pwd`:/workdir \
   -w /workdir \
-  ghcr.io/pinto0309/onnx2tf:1.26.1
+  ghcr.io/pinto0309/onnx2tf:1.26.3
 
   or
 
@@ -307,7 +307,7 @@ Video speed is adjusted approximately 50 times slower than actual speed.
   docker run --rm -it \
   -v `pwd`:/workdir \
   -w /workdir \
-  docker.io/pinto0309/onnx2tf:1.26.1
+  docker.io/pinto0309/onnx2tf:1.26.3
 
   or
 
@@ -1448,9 +1448,13 @@ For example, take a model with multiple inputs and multiple outputs as shown in
 When converting to TensorFlow.js, process as follows.
 
 ```bash
-pip install tensorflowjs
+pip install -U --no-deps \
+tensorflowjs \
+tensorflow_decision_forests \
+ydf \
+tensorflow_hub
 
-onnx2tf -i mobilenetv2-12.onnx -ois input:1,3,224,224 -osd
+onnx2tf -i mobilenetv2-12.onnx -ois input:1,3,224,224 -osd -dgc
 
 tensorflowjs_converter \
 --input_format tf_saved_model \

{onnx2tf-1.26.1 → onnx2tf-1.26.3}/onnx2tf/__init__.py

@@ -1,3 +1,3 @@
 from onnx2tf.onnx2tf import convert, main
 
-__version__ = '1.26.1'
+__version__ = '1.26.3'

{onnx2tf-1.26.1 → onnx2tf-1.26.3}/onnx2tf/onnx2tf.py

@@ -1626,6 +1626,7 @@ def convert(
                             mean,
                             std,
                         ]
+
             elif custom_input_op_name_np_data_path is not None:
                 for param in custom_input_op_name_np_data_path:
                     if len(param) != 4:
@@ -1652,11 +1653,14 @@ def convert(
 
             # representative_dataset_gen
             def representative_dataset_gen():
-                for idx in range(data_count):
+                batch_size = model.inputs[0].shape[0]
+                if not isinstance(batch_size, int):
+                    batch_size = 1
+                for idx in range(0, data_count, batch_size):
                     yield_data_dict = {}
                     for model_input_name in model_input_name_list:
                         calib_data, mean, std = calib_data_dict[model_input_name]
-                        normalized_calib_data: np.ndarray = (calib_data[idx] - mean) / std
+                        normalized_calib_data: np.ndarray = (calib_data[idx:idx+batch_size] - mean) / std
                         yield_data_dict[model_input_name] = tf.cast(tf.convert_to_tensor(normalized_calib_data), tf.float32)
                     yield yield_data_dict
 
@@ -1708,7 +1712,7 @@ def convert(
                     inf_type_input = tf.float32
                 else:
                     inf_type_input = tf.int8
-                
+
                 if output_quant_dtype == 'int8':
                     inf_type_output = tf.int8
                 elif output_quant_dtype == 'uint8':

{onnx2tf-1.26.1 → onnx2tf-1.26.3}/onnx2tf/ops/MatMul.py

@@ -87,7 +87,7 @@ def make_node(
     input_tensor_2_is_one_d = False
     if input_tensor_1.shape is not None \
         and len(input_tensor_1.shape) == 1:
-        input_tensor_1 = tf.expand_dims(input_tensor_2, axis=0)
+        input_tensor_1 = tf.expand_dims(input_tensor_1, axis=0)
         input_tensor_1_is_one_d = True
     elif input_tensor_2.shape is not None \
             and len(input_tensor_2.shape) == 1:

{onnx2tf-1.26.1 → onnx2tf-1.26.3/onnx2tf.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: onnx2tf
-Version: 1.26.1
+Version: 1.26.3
 Summary: Self-Created Tools to convert ONNX files (NCHW) to TensorFlow/TFLite/Keras format (NHWC). The purpose of this tool is to solve the massive Transpose extrapolation problem in onnx-tensorflow (onnx-tf).
 Home-page: https://github.com/PINTO0309/onnx2tf
 Author: Katsuya Hyodo
@@ -314,7 +314,7 @@ Video speed is adjusted approximately 50 times slower than actual speed.
   docker run --rm -it \
   -v `pwd`:/workdir \
   -w /workdir \
-  ghcr.io/pinto0309/onnx2tf:1.26.1
+  ghcr.io/pinto0309/onnx2tf:1.26.3
 
   or
 
@@ -322,7 +322,7 @@ Video speed is adjusted approximately 50 times slower than actual speed.
   docker run --rm -it \
   -v `pwd`:/workdir \
   -w /workdir \
-  docker.io/pinto0309/onnx2tf:1.26.1
+  docker.io/pinto0309/onnx2tf:1.26.3
 
   or
 
@@ -1463,9 +1463,13 @@ For example, take a model with multiple inputs and multiple outputs as shown in
 When converting to TensorFlow.js, process as follows.
 
 ```bash
-pip install tensorflowjs
+pip install -U --no-deps \
+tensorflowjs \
+tensorflow_decision_forests \
+ydf \
+tensorflow_hub
 
-onnx2tf -i mobilenetv2-12.onnx -ois input:1,3,224,224 -osd
+onnx2tf -i mobilenetv2-12.onnx -ois input:1,3,224,224 -osd -dgc
 
 tensorflowjs_converter \
 --input_format tf_saved_model \