| 🟡 |
bioimageio format validation |
| status |
valid-format |
| source |
https://hypha.aicell.io/bioimage-io/artifacts/sympathetic-mosquito/files/rdf.yaml?version=v0 |
| id |
sympathetic-mosquito |
| format version |
model 0.4.10 |
| bioimageio.spec |
0.5.6.0 |
|
Location |
Details |
| ✔️ |
|
Successfully created `ModelDescr` instance. |
| ✔️ |
|
bioimageio.spec format validation model 0.4.10 |
| ❌ |
weights.keras_hdf5 |
Reproduce test outputs from test inputs (keras_hdf5) |
| ❌ |
weights.keras_hdf5 |
Error when deserializing class 'Conv2D' using config={'name': 'conv2d', 'trainable': True, 'dtype': 'float32', 'filters': 64, 'kernel_size': [3, 3], 'strides': [1, 1], 'padding': 'same', 'data_format': 'channels_last', 'dilation_rate': [1, 1], 'activation': 'relu', 'use_bias': True, 'kernel_initializer': {'class_name': 'VarianceScaling', 'config': {'scale': 2.0, 'mode': 'fan_in', 'distribution': 'truncated_normal', 'seed': None, 'dtype': 'float32'}}, 'bias_initializer': {'class_name': 'Zeros', 'config': {'dtype': 'float32'}}, 'kernel_regularizer': None, 'bias_regularizer': None, 'activity_regularizer': None, 'kernel_constraint': None, 'bias_constraint': None}.
Exception encountered: could not be deserialized properly. Please ensure that components that are Python object instances (layers, models, etc.) returned by `get_config()` are explicitly deserialized in the model's `from_config()` method.
config={'module': 'keras.initializers', 'class_name': 'VarianceScaling', 'config': {'scale': 2.0, 'mode': 'fan_in', 'distribution': 'truncated_normal', 'seed': None, 'dtype': 'float32'}, 'registered_name': None}.
Exception encountered: VarianceScaling.__init__() got an unexpected keyword argument 'dtype' |
|
|
See Traceback 1. |
|
weights.keras_hdf5 |
recommended conda environment (Reproduce test outputs from test inputs (keras_hdf5)) |
|
|
%YAML 1.2
---
channels:
- conda-forge
- nodefaults
dependencies:
- bioimageio.core
- pip
- tensorflow ==2.17
|
|
weights.keras_hdf5 |
conda compare (Reproduce test outputs from test inputs (keras_hdf5)) |
|
|
Success. All the packages in the specification file are present in the environment with matching
version and build string. |
Traceback 1
╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/saving/serialization_lib.py:733 in de │
│ │
│ 730 │ safe_mode_scope = SafeModeScope(safe_mode) │
│ 731 │ with custom_obj_scope, safe_mode_scope: │
│ 732 │ │ try: │
│ ❱ 733 │ │ │ instance = cls.from_config(inner_config) │
│ 734 │ │ except TypeError as e: │
│ 735 │ │ │ raise TypeError( │
│ 736 │ │ │ │ f"{cls} could not be deserialized properly. Please" │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/initializers/initializer.py:81 in fro │
│ │
│ 78 │ │ Returns: │
│ 79 │ │ │ An `Initializer` instance. │
│ 80 │ │ """ │
│ ❱ 81 │ │ return cls(**config) │
│ 82 │ │
│ 83 │ def clone(self): │
│ 84 │ │ return self.__class__.from_config(self.get_config()) │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
TypeError: VarianceScaling.__init__() got an unexpected keyword argument 'dtype'
During handling of the above exception, another exception occurred:
╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/ops/operation.py:317 in from_config │
│ │
│ 314 │ │ │ │ policy = policy.name │
│ 315 │ │ │ config["dtype"] = policy │
│ 316 │ │ try: │
│ ❱ 317 │ │ │ return cls(**config) │
│ 318 │ │ except Exception as e: │
│ 319 │ │ │ raise TypeError( │
│ 320 │ │ │ │ f"Error when deserializing class '{cls.__name__}' using " │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/layers/convolutional/conv2d.py:118 in │
│ │
│ 115 │ │ bias_constraint=None, │
│ 116 │ │ **kwargs, │
│ 117 │ ): │
│ ❱ 118 │ │ super().__init__( │
│ 119 │ │ │ rank=2, │
│ 120 │ │ │ filters=filters, │
│ 121 │ │ │ kernel_size=kernel_size, │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/layers/convolutional/base_conv.py:126 │
│ │
│ 123 │ │ self.data_format = standardize_data_format(data_format) │
│ 124 │ │ self.activation = activations.get(activation) │
│ 125 │ │ self.use_bias = use_bias │
│ ❱ 126 │ │ self.kernel_initializer = initializers.get(kernel_initializer) │
│ 127 │ │ self.bias_initializer = initializers.get(bias_initializer) │
│ 128 │ │ self.kernel_regularizer = regularizers.get(kernel_regularizer) │
│ 129 │ │ self.bias_regularizer = regularizers.get(bias_regularizer) │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/initializers/__init__.py:133 in get │
│ │
│ 130 │ if identifier is None: │
│ 131 │ │ return None │
│ 132 │ if isinstance(identifier, dict): │
│ ❱ 133 │ │ obj = deserialize(identifier) │
│ 134 │ elif isinstance(identifier, str): │
│ 135 │ │ config = {"class_name": str(identifier), "config": {}} │
│ 136 │ │ obj = deserialize(config) │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/initializers/__init__.py:75 in deseri │
│ │
│ 72 @keras_export("keras.initializers.deserialize") │
│ 73 def deserialize(config, custom_objects=None): │
│ 74 │ """Returns a Keras initializer object via its configuration.""" │
│ ❱ 75 │ return serialization_lib.deserialize_keras_object( │
│ 76 │ │ config, │
│ 77 │ │ module_objects=ALL_OBJECTS_DICT, │
│ 78 │ │ custom_objects=custom_objects, │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/saving/serialization_lib.py:596 in de │
│ │
│ 593 │ │ │ │ │ ), │
│ 594 │ │ │ │ │ custom_objects=custom_objects, │
│ 595 │ │ │ │ ) │
│ ❱ 596 │ │ │ return deserialize_keras_object( │
│ 597 │ │ │ │ serialize_with_public_class( │
│ 598 │ │ │ │ │ module_objects[config], inner_config=inner_config │
│ 599 │ │ │ │ ), │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/saving/serialization_lib.py:735 in de │
│ │
│ 732 │ │ try: │
│ 733 │ │ │ instance = cls.from_config(inner_config) │
│ 734 │ │ except TypeError as e: │
│ ❱ 735 │ │ │ raise TypeError( │
│ 736 │ │ │ │ f"{cls} could not be deserialized properly. Please" │
│ 737 │ │ │ │ " ensure that components that are Python object" │
│ 738 │ │ │ │ " instances (layers, models, etc.) returned by" │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
TypeError: <class 'keras.src.initializers.random_initializers.VarianceScaling'> could not be deserialized properly. Please ensure that components that are Python object instances (layers, models, etc.) returned by `get_config()` are explicitly deserialized in the model's `from_config()` method.
config={'module': 'keras.initializers', 'class_name': 'VarianceScaling', 'config': {'scale': 2.0, 'mode': 'fan_in', 'distribution': 'truncated_normal', 'seed': None, 'dtype': 'float32'}, 'registered_name': None}.
Exception encountered: VarianceScaling.__init__() got an unexpected keyword argument 'dtype'
During handling of the above exception, another exception occurred:
╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮
│ /home/runner/.local/lib/python3.12/site-packages/bioimageio/core/_resource_tests.py:784 in _test │
│ │
│ 781 │ │ test_input = get_test_input_sample(model) │
│ 782 │ │ expected = get_test_output_sample(model) │
│ 783 │ │ │
│ ❱ 784 │ │ with create_prediction_pipeline( │
│ 785 │ │ │ bioimageio_model=model, devices=devices, weight_format=weight_format │
│ 786 │ │ ) as prediction_pipeline: │
│ 787 │ │ │ results = prediction_pipeline.predict_sample_without_blocking(test_input) │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/bioimageio/core/_prediction_pipeline.py:371 in │
│ │
│ 368 │ │ │ f"deprecated create_prediction_pipeline kwargs: {set(deprecated_kwargs)}" │
│ 369 │ │ ) │
│ 370 │ │
│ ❱ 371 │ model_adapter = model_adapter or create_model_adapter( │
│ 372 │ │ model_description=bioimageio_model, │
│ 373 │ │ devices=devices, │
│ 374 │ │ weight_format_priority_order=weights_format and (weights_format,), │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/bioimageio/core/backends/_model_adapter.py:169 │
│ │
│ 166 │ │ assert errors │
│ 167 │ │ if len(weight_format_priority_order) == 1: │
│ 168 │ │ │ assert len(errors) == 1 │
│ ❱ 169 │ │ │ raise errors[0] │
│ 170 │ │ │
│ 171 │ │ else: │
│ 172 │ │ │ msg = ( │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/bioimageio/core/backends/_model_adapter.py:158 │
│ │
│ 155 │ │ │ │ │ except Exception: │
│ 156 │ │ │ │ │ │ from .tensorflow_backend import KerasModelAdapter │
│ 157 │ │ │ │ │ │
│ ❱ 158 │ │ │ │ │ return KerasModelAdapter( │
│ 159 │ │ │ │ │ │ model_description=model_description, devices=devices │
│ 160 │ │ │ │ │ ) │
│ 161 │ │ │ │ except Exception as e: │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/bioimageio/core/backends/keras_backend.py:87 in │
│ │
│ 84 │ │ │ │ with temp_path.open("wb") as f: │
│ 85 │ │ │ │ │ shutil.copyfileobj(weight_reader, f) │
│ 86 │ │ │ │ │
│ ❱ 87 │ │ │ │ self._network = keras.models.load_model(temp_path) │
│ 88 │ │ │
│ 89 │ │ self._output_axes = [ │
│ 90 │ │ │ tuple(a.id for a in get_axes_infos(out)) │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/saving/saving_api.py:196 in load_mode │
│ │
│ 193 │ │ │ safe_mode=safe_mode, │
│ 194 │ │ ) │
│ 195 │ if str(filepath).endswith((".h5", ".hdf5")): │
│ ❱ 196 │ │ return legacy_h5_format.load_model_from_hdf5( │
│ 197 │ │ │ filepath, │
│ 198 │ │ │ custom_objects=custom_objects, │
│ 199 │ │ │ compile=compile, │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/legacy/saving/legacy_h5_format.py:137 │
│ │
│ 134 │ │ legacy_scope = saving_options.keras_option_scope(use_legacy_config=True) │
│ 135 │ │ safe_mode_scope = serialization_lib.SafeModeScope(safe_mode) │
│ 136 │ │ with legacy_scope, safe_mode_scope: │
│ ❱ 137 │ │ │ model = saving_utils.model_from_config( │
│ 138 │ │ │ │ model_config, custom_objects=custom_objects │
│ 139 │ │ │ ) │
│ 140 │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/legacy/saving/saving_utils.py:83 in m │
│ │
│ 80 │ │ │ function_dict["config"]["closure"] = function_config[2] │
│ 81 │ │ │ config["config"]["function"] = function_dict │
│ 82 │ │
│ ❱ 83 │ return serialization.deserialize_keras_object( │
│ 84 │ │ config, │
│ 85 │ │ module_objects=MODULE_OBJECTS.ALL_OBJECTS, │
│ 86 │ │ custom_objects=custom_objects, │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/legacy/saving/serialization.py:488 in │
│ │
│ 485 │ │ │ custom_objects = custom_objects or {} │
│ 486 │ │ │ │
│ 487 │ │ │ if "custom_objects" in arg_spec.args: │
│ ❱ 488 │ │ │ │ deserialized_obj = cls.from_config( │
│ 489 │ │ │ │ │ cls_config, │
│ 490 │ │ │ │ │ custom_objects={ │
│ 491 │ │ │ │ │ │ **object_registration.GLOBAL_CUSTOM_OBJECTS, │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/models/model.py:694 in from_config │
│ │
│ 691 │ │ │ # (but not Functional subclasses with a custom __init__) │
│ 692 │ │ │ from keras.src.models.functional import functional_from_config │
│ 693 │ │ │ │
│ ❱ 694 │ │ │ return functional_from_config( │
│ 695 │ │ │ │ cls, config, custom_objects=custom_objects │
│ 696 │ │ │ ) │
│ 697 │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/models/functional.py:558 in functiona │
│ │
│ 555 │ │
│ 556 │ # First, we create all layers and enqueue nodes to be processed │
│ 557 │ for layer_data in functional_config["layers"]: │
│ ❱ 558 │ │ process_layer(layer_data) │
│ 559 │ │
│ 560 │ # Then we process nodes in order of layer depth. │
│ 561 │ # Nodes that cannot yet be processed (if the inbound node │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/models/functional.py:521 in process_l │
│ │
│ 518 │ │ if "module" not in layer_data: │
│ 519 │ │ │ # Legacy format deserialization (no "module" key) │
│ 520 │ │ │ # used for H5 and SavedModel formats │
│ ❱ 521 │ │ │ layer = saving_utils.model_from_config( │
│ 522 │ │ │ │ layer_data, custom_objects=custom_objects │
│ 523 │ │ │ ) │
│ 524 │ │ else: │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/legacy/saving/saving_utils.py:83 in m │
│ │
│ 80 │ │ │ function_dict["config"]["closure"] = function_config[2] │
│ 81 │ │ │ config["config"]["function"] = function_dict │
│ 82 │ │
│ ❱ 83 │ return serialization.deserialize_keras_object( │
│ 84 │ │ config, │
│ 85 │ │ module_objects=MODULE_OBJECTS.ALL_OBJECTS, │
│ 86 │ │ custom_objects=custom_objects, │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/legacy/saving/serialization.py:497 in │
│ │
│ 494 │ │ │ │ ) │
│ 495 │ │ │ else: │
│ 496 │ │ │ │ with object_registration.CustomObjectScope(custom_objects): │
│ ❱ 497 │ │ │ │ │ deserialized_obj = cls.from_config(cls_config) │
│ 498 │ │ else: │
│ 499 │ │ │ # Then `cls` may be a function returning a class. │
│ 500 │ │ │ # in this case by convention `config` holds │
│ │
│ /home/runner/.local/lib/python3.12/site-packages/keras/src/ops/operation.py:319 in from_config │
│ │
│ 316 │ │ try: │
│ 317 │ │ │ return cls(**config) │
│ 318 │ │ except Exception as e: │
│ ❱ 319 │ │ │ raise TypeError( │
│ 320 │ │ │ │ f"Error when deserializing class '{cls.__name__}' using " │
│ 321 │ │ │ │ f"config={config}.\n\nException encountered: {e}" │
│ 322 │ │ │ ) │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
TypeError: Error when deserializing class 'Conv2D' using config={'name': 'conv2d', 'trainable': True, 'dtype': 'float32', 'filters': 64, 'kernel_size': [3, 3], 'strides': [1, 1], 'padding': 'same', 'data_format': 'channels_last', 'dilation_rate': [1, 1], 'activation': 'relu', 'use_bias': True, 'kernel_initializer': {'class_name': 'VarianceScaling', 'config': {'scale': 2.0, 'mode': 'fan_in', 'distribution': 'truncated_normal', 'seed': None, 'dtype': 'float32'}}, 'bias_initializer': {'class_name': 'Zeros', 'config': {'dtype': 'float32'}}, 'kernel_regularizer': None, 'bias_regularizer': None, 'activity_regularizer': None, 'kernel_constraint': None, 'bias_constraint': None}.
Exception encountered: <class 'keras.src.initializers.random_initializers.VarianceScaling'> could not be deserialized properly. Please ensure that components that are Python object instances (layers, models, etc.) returned by `get_config()` are explicitly deserialized in the model's `from_config()` method.
config={'module': 'keras.initializers', 'class_name': 'VarianceScaling', 'config': {'scale': 2.0, 'mode': 'fan_in', 'distribution': 'truncated_normal', 'seed': None, 'dtype': 'float32'}, 'registered_name': None}.
Exception encountered: VarianceScaling.__init__() got an unexpected keyword argument 'dtype'