Enable tf2onnx installation

Signed-off-by: Tom Wildenhain <[email protected]>

.azure-pipelines/linux-conda-CI-tf-keras.yml

@@ -66,6 +66,7 @@ jobs:
       pip install -r requirements.txt
       pip install -r requirements-dev.txt
       $(INSTALL_ORT)
+      pip install git+https://github.com/onnx/tensorflow-onnx
     displayName: 'Install dependencies'
 
   - script: |
@@ -84,129 +85,3 @@ jobs:
       testResultsFiles: '**/test-results.xml'
       testRunTitle: 'Python $(python.version)'
     condition: succeededOrFailed()
-
-- job: 'Tests_Nightly'
-  timeoutInMinutes: 180
-  pool:
-    vmImage: 'Ubuntu-16.04'
-  strategy:
-    matrix:
-      Python36-onnx1.2:
-        python.version: '3.6'
-        ONNX_PATH: onnx==1.2.3
-        INSTALL_KERAS: pip install keras==2.1.6
-        UNINSTALL_KERAS:
-        INSTALL_TENSORFLOW: pip install tensorflow==1.15.0
-        INSTALL_ORT: pip install onnxruntime==1.1.1
-        INSTALL_KERAS_RESNET:
-        INSTALL_TRANSFORMERS:
-        NIGHTLY_BUILD_TEST: python run_all.py --exclude "test_keras_applications_v2.py"
-
-      Python36-onnx1.5:
-        python.version: '3.6'
-        ONNX_PATH: onnx==1.5.0
-        INSTALL_KERAS: pip install keras==2.2.4
-        UNINSTALL_KERAS:
-        INSTALL_TENSORFLOW: pip install tensorflow==1.15.0
-        INSTALL_ORT: pip install onnxruntime==1.1.1
-        INSTALL_KERAS_RESNET: pip install keras-resnet
-        INSTALL_TRANSFORMERS:
-        NIGHTLY_BUILD_TEST: python run_all.py --exclude "test_keras_applications_v2.py"
-
-      Python37:
-        python.version: '3.7.3'
-        ONNX_PATH: onnx==1.8.0
-        INSTALL_KERAS: pip install keras==2.3.1
-        UNINSTALL_KERAS:
-        INSTALL_TENSORFLOW: pip install tensorflow==1.15.0
-        INSTALL_ORT: pip install -i https://test.pypi.org/simple/ ort-nightly
-        INSTALL_KERAS_RESNET: pip install keras-resnet
-        INSTALL_TRANSFORMERS:
-        NIGHTLY_BUILD_TEST: python run_all.py --exclude "test_keras_applications_v2.py"
-
-      Python37-official-ort:
-        python.version: '3.7.3'
-        ONNX_PATH: onnx==1.6.0
-        INSTALL_KERAS: pip install keras==2.3.1
-        UNINSTALL_KERAS:
-        INSTALL_TENSORFLOW: pip install tensorflow==1.15.0
-        INSTALL_ORT: pip install onnxruntime==1.3.0
-        INSTALL_KERAS_RESNET: pip install keras-resnet
-        INSTALL_TRANSFORMERS:
-        NIGHTLY_BUILD_TEST: python run_all.py --exclude "test_keras_applications_v2.py"
-
-      Python38-tf2:
-        python.version: '3.8'
-        ONNX_PATH: onnx==1.8.0
-        INSTALL_KERAS:
-        UNINSTALL_KERAS: pip uninstall keras -y
-        INSTALL_TENSORFLOW: pip install tensorflow==2.2.0
-        INSTALL_ORT: pip install -i https://test.pypi.org/simple/ ort-nightly
-        INSTALL_KERAS_RESNET: pip install keras-resnet
-        INSTALL_TRANSFORMERS: pip install transformers==3.4.0
-        NIGHTLY_BUILD_TEST: python run_all_v2.py
-
-    maxParallel: 3
-
-  steps:
-  - script: sudo install -d -m 0777 /home/vsts/.conda/envs
-    displayName: Fix Conda permissions
-
-  - task: CondaEnvironment@1
-    inputs:
-      createCustomEnvironment: true
-      environmentName: 'py$(python.version)'
-      packageSpecs: 'python=$(python.version)'
-
-  - script: |
-      python -m pip install --upgrade pip
-      conda config --set always_yes yes --set changeps1 no
-      pip install $(ONNX_PATH)
-      pip install h5py==2.9.0
-      $(INSTALL_TENSORFLOW)
-      $(INSTALL_KERAS)
-      pip install git+https://github.com/microsoft/onnxconverter-common
-      pip install -r requirements.txt
-      pip install -r requirements-dev.txt
-      $(INSTALL_ORT)
-      pip install opencv-python
-      pip install tqdm
-      pip install keras-segmentation==0.2.0
-      git clone https://github.com/matterport/Mask_RCNN
-      cd Mask_RCNN
-      pip install -r requirements.txt
-      python setup.py install
-      cd ..
-      pip install matplotlib
-      git clone https://github.com/qqwweee/keras-yolo3
-      $(INSTALL_KERAS_RESNET)
-      pip install git+https://www.github.com/keras-team/keras-contrib.git
-      pip install keras-tcn==2.8.3
-      $(UNINSTALL_KERAS)
-      pip install git+https://github.com/qubvel/efficientnet
-      $(INSTALL_TRANSFORMERS)
-      pip install keras-self-attention
-      pip install git+https://github.com/onnx/tensorflow-onnx
-    displayName: 'Install dependencies'
-
-  - script: |
-      pip install -e .
-      python -c "import onnxruntime"
-      pytest tests --doctest-modules --junitxml=junit/test-results.xml
-      cd applications/nightly_build
-      $(NIGHTLY_BUILD_TEST)
-    displayName: 'pytest'
-
-  - script: |
-      pip install -e .
-      python -c "import onnxruntime"
-      coverage run --include=keras2onnx/* -m pytest tests/test_layers.py
-      coverage report -m
-      coverage html
-    displayName: 'coverage'
-
-  - task: PublishTestResults@2
-    inputs:
-      testResultsFiles: '**/test-results-*.xml'
-      testRunTitle: 'Python $(python.version)'
-    condition: succeededOrFailed()

.azure-pipelines/linux-conda-CI.yml

@@ -73,6 +73,7 @@ jobs:
       pip install -r requirements.txt
       pip install -r requirements-dev.txt
       $(INSTALL_ORT)
+      pip install git+https://github.com/onnx/tensorflow-onnx
     displayName: 'Install dependencies'
 
   - script: |

.azure-pipelines/win32-conda-CI-tf-keras.yml

@@ -69,6 +69,7 @@ jobs:
       pip install -r requirements.txt
       pip install -r requirements-dev.txt
       %INSTALL_ORT%
+      pip install git+https://github.com/onnx/tensorflow-onnx
     displayName: 'Install dependencies'
 
   - script: |

.azure-pipelines/win32-conda-CI.yml

@@ -69,6 +69,7 @@ jobs:
       pip install -r requirements.txt
       pip install -r requirements-dev.txt
       %INSTALL_ORT%
+      pip install git+https://github.com/onnx/tensorflow-onnx
     displayName: 'Install dependencies'
 
   - script: |

keras2onnx/_tf_utils.py

@@ -85,8 +85,8 @@ def tf_attrs_to_onnx(node):
     return attrs
 
 
-def to_tf_tensor_spec(onnx_type, name=None):
-    shp = [1 if isinstance(n_, str) else n_ for n_ in onnx_type.shape]
+def to_tf_tensor_spec(onnx_type, name=None, unknown_dim=1):
+    shp = [unknown_dim if isinstance(n_, str) else n_ for n_ in onnx_type.shape]
     return tensorflow.TensorSpec(shp,
                                  mapping.TENSOR_TYPE_TO_NP_TYPE[onnx_type.to_onnx_type().tensor_type.elem_type],
                                  name=name)

keras2onnx/main.py

@@ -18,7 +18,7 @@
 from ._parser_tf import build_layer_output_from_model
 
 
-def _process_initial_types(initial_types):
+def _process_initial_types(initial_types, unknown_dim=1):
     if initial_types is None:
         return None
 
@@ -30,7 +30,7 @@ def _process_initial_types(initial_types):
         if isinstance(initial_types[c_], str):
             name = initial_types[c_]
             type_idx = c_ + 1
-        ts_spec = to_tf_tensor_spec(initial_types[type_idx], name)
+        ts_spec = to_tf_tensor_spec(initial_types[type_idx], name, unknown_dim)
         input_specs.append(ts_spec)
         c_ += 1 if name is None else 2
 
@@ -41,10 +41,10 @@ def convert_keras_tf2onnx(model, name=None, doc_string='', target_opset=None, in
                           channel_first_inputs=None, debug_mode=False, custom_op_conversions=None):
     if target_opset is None:
         target_opset = 13
-    input_signature = _process_initial_types(initial_types)
+    input_signature = _process_initial_types(initial_types, unknown_dim=None)
 
     import tf2onnx
-    model, external_tensor_storage = tf2onnx.convert.from_keras(model, input_signature, opset=target_opset)
+    model, external_tensor_storage = tf2onnx.convert.from_keras(model, input_signature, opset=target_opset, inputs_as_nchw=channel_first_inputs)
 
     return model
 

tests/test_utils.py

@@ -211,6 +211,8 @@ def run_onnx_runtime(case_name, onnx_model, data, expected, model_files, rtol=1.
     elif not isinstance(expected, list):
         expected = [expected]
 
+    for n_ in range(len(expected)):
+        np.testing.assert_allclose(expected[n_], actual[n_], rtol=rtol, atol=atol)
     res = all(np.allclose(expected[n_], actual[n_], rtol=rtol, atol=atol) for n_ in range(len(expected)))
 
     if res and temp_model_file not in model_files:  # still keep the failed case files for the diagnosis.