Loading up Json_files built and trained in Keras 2 for Keras 3

[manuelpaeza]

Using Keras 3, I am trying to load up a built and trained model from Keras 2 API that is stored in .json with weights stored in .h5. The model file is the following: cnn_model.json. Since model_from_json does not exist in Keras 3, I rewrote the function from the Keras 2 API so that I can load the .json file. With Keras 3 (with torch backend), I am trying to load the model and the weights with the following code

import os
import keras 
import json 

os.environ["KERAS_BACKEND"] = "torch"

def model_from_json(json_string, custom_objects=None):
    """Parses a JSON model configuration string and returns a model instance.

    Args:
        json_string: JSON string encoding a model configuration.
        custom_objects: Optional dictionary mapping names
            (strings) to custom classes or functions to be
            considered during deserialization.

    Returns:
        A Keras model instance (uncompiled).

    model_config = json.loads(json_string)
    return deserialize_keras_object(model_config, custom_objects=custom_objects) 

def model_torch():
    model_name = 'cnn_model' #model file name 
    model_file = model_name + '.json'
    with open(model_file, 'r') as json_file:
        print('USING MODEL:' + model_file)
        loaded_model_json = json_file.read()

    loaded_model = model_from_json(loaded_model_json)
    loaded_model.load_weights(model_name + '.h5')
    loaded_model.compile('sgd', 'mse')

if __name__ == "__main__":
    model_torch()

However, when I run this code, I obtain the error below (as shown below). With this, I have the three following questions:

1. How does one possibly fix this error given that the model I want to load (in Keras 3) was built and trained in tensorflow-keras 2?
2. Is it better to rebuild the model in Keras using the load_model() function in Keras 3, and if so, how can you translate the weights from the .h5 file that was created in tensorflow-keras 2 to keras 3?
3. To rebuild how, how should one translate the json dictionary to actual code?

Error I obtain: TypeError: Could not locate class 'Sequential'. Make sure custom classes are decorated with@keras.saving.register_keras_serializable(). Full object config: {'class_name': 'Sequential', 'config': {'name': 'sequential', 'layers': [{'class_name': 'Conv2D', 'config': {'name': 'conv2d_20', 'trainable': True, 'batch_input_shape': [None, 50, 50, 1], 'dtype': 'float32', 'filters': 32, 'kernel_size': [3, 3], 'strides': [1, 1], 'padding': 'valid', 'data_format': 'channels_last', 'dilation_rate': [1, 1], 'activation': 'relu', 'use_bias': True, 'kernel_initializer': {'class_name': 'VarianceScaling', 'config': {'scale': 1.0, 'mode': 'fan_avg', 'distribution': 'uniform', '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}}, {'class_name': 'Activation', 'config': {'name': 'activation_13', 'trainable': True, 'dtype': 'float32', 'activation': 'relu'}}, {'class_name': 'Conv2D', 'config': {'name': 'conv2d_21', 'trainable': True, 'dtype': 'float32', 'filters': 32, 'kernel_size': [3, 3], 'strides': [1, 1], 'padding': 'valid', 'data_format': 'channels_last', 'dilation_rate': [1, 1], 'activation': 'linear', 'use_bias': True, 'kernel_initializer': {'class_name': 'VarianceScaling', 'config': {'scale': 1.0, 'mode': 'fan_avg', 'distribution': 'uniform', '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}}, {'class_name': 'Activation', 'config': {'name': 'activation_14', 'trainable': True, 'dtype': 'float32', 'activation': 'relu'}}, {'class_name': 'MaxPooling2D', 'config': {'name': 'max_pooling2d_10', 'trainable': True, 'dtype': 'float32', 'pool_size': [2, 2], 'padding': 'valid', 'strides': [2, 2], 'data_format': 'channels_last'}}, {'class_name': 'Dropout', 'config': {'name': 'dropout_17', 'trainable': True, 'dtype': 'float32', 'rate': 0.25, 'noise_shape': None, 'seed': None}}, {'class_name': 'Conv2D', 'config': {'name': 'conv2d_22', 'trainable': True, 'dtype': 'float32', 'filters': 64, 'kernel_size': [3, 3], 'strides': [1, 1], 'padding': 'same', 'data_format': 'channels_last', 'dilation_rate': [1, 1], 'activation': 'linear', 'use_bias': True, 'kernel_initializer': {'class_name': 'VarianceScaling', 'config': {'scale': 1.0, 'mode': 'fan_avg', 'distribution': 'uniform', '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}}, {'class_name': 'Activation', 'config': {'name': 'activation_15', 'trainable': True, 'dtype': 'float32', 'activation': 'relu'}}, {'class_name': 'Conv2D', 'config': {'name': 'conv2d_23', 'trainable': True, 'dtype': 'float32', 'filters': 64, 'kernel_size': [3, 3], 'strides': [1, 1], 'padding': 'valid', 'data_format': 'channels_last', 'dilation_rate': [1, 1], 'activation': 'linear', 'use_bias': True, 'kernel_initializer': {'class_name': 'VarianceScaling', 'config': {'scale': 1.0, 'mode': 'fan_avg', 'distribution': 'uniform', '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}}, {'class_name': 'Activation', 'config': {'name': 'activation_16', 'trainable': True, 'dtype': 'float32', 'activation': 'relu'}}, {'class_name': 'MaxPooling2D', 'config': {'name': 'max_pooling2d_11', 'trainable': True, 'dtype': 'float32', 'pool_size': [2, 2], 'padding': 'valid', 'strides': [2, 2], 'data_format': 'channels_last'}}, {'class_name': 'Dropout', 'config': {'name': 'dropout_18', 'trainable': True, 'dtype': 'float32', 'rate': 0.25, 'noise_shape': None, 'seed': None}}, {'class_name': 'Flatten', 'config': {'name': 'flatten_8', 'trainable': True, 'dtype': 'float32', 'data_format': 'channels_last'}}, {'class_name': 'Dense', 'config': {'name': 'dense_15', 'trainable': True, 'dtype': 'float32', 'units': 512, 'activation': 'linear', 'use_bias': True, 'kernel_initializer': {'class_name': 'VarianceScaling', 'config': {'scale': 1.0, 'mode': 'fan_avg', 'distribution': 'uniform', '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}}, {'class_name': 'Activation', 'config': {'name': 'activation_17', 'trainable': True, 'dtype': 'float32', 'activation': 'relu'}}, {'class_name': 'Dropout', 'config': {'name': 'dropout_19', 'trainable': True, 'dtype': 'float32', 'rate': 0.5, 'noise_shape': None, 'seed': None}}, {'class_name': 'Dense', 'config': {'name': 'dense_16', 'trainable': True, 'dtype': 'float32', 'units': 2, 'activation': 'linear', 'use_bias': True, 'kernel_initializer': {'class_name': 'VarianceScaling', 'config': {'scale': 1.0, 'mode': 'fan_avg', 'distribution': 'uniform', '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}}, {'class_name': 'Activation', 'config': {'name': 'activation_18', 'trainable': True, 'dtype': 'float32', 'activation': 'softmax'}}]}, 'keras_version': '2.2.4-tf', 'backend': 'tensorflow'}

[sachinprasadhs]

Hi, Cross version saving and loading is not guaranteed to work between Keras 2 and 3. It's suggested to migrate your code to Keras and and then use the save and load method.

[manuelpaeza]

Hello, then I have this dilemma. The models I have were trained with a older version of keras, tensorflow, and python. However, I do not have the training file for the models. I do have is the model in .json format and the weights in .h5 format. I have the following code that I used to load up the model in Tensorflow 2.15.0 and recreate it using .keras file extension to use in Keras 3

#!/usr/bin/env python

import numpy as np
import os

from caiman.paths import caiman_datadir

os.environ["KERAS_BACKEND"] = "tensorflow"
from tensorflow.keras.models import model_from_json
use_keras = True

def recreate_model():
    os.environ['CUDA_VISIBLE_DEVICES'] = '-1'

    model_name = os.path.join(caiman_datadir(), 'model', 'cnn_model_online')
    if use_keras:
        model_file = model_name + ".json"
        with open(model_file, 'r') as json_file:
            print('USING MODEL:' + model_file)
            loaded_model_json = json_file.read()

        loaded_model = model_from_json(loaded_model_json)
        loaded_model.load_weights(model_name + '.h5')
        loaded_model.save(model_name + '.keras')


if __name__ == "__main__":
    recreate_model()

However, when I try to load up the .keras model in Keras 3.4.1 with the following code:

#!/usr/bin/env python

import os
import keras 
os.environ["KERAS_BACKEND"] = "torch"
from keras.models import load_model 
use_keras = True

from caiman.paths import caiman_datadir

def test_torch():
    os.environ['CUDA_VISIBLE_DEVICES'] = '-1'

    try:
        model_name = os.path.join(caiman_datadir(), 'model', 'cnn_model')
        if use_keras:
            model_file = model_name + ".keras"
            loaded_model = load_model(model_file, compile=False)
            loaded_model.compile(optimizer='sgd', loss='mse')
    except:
        raise Exception(f'NN model could not be loaded. use_keras = {use_keras}')

if __name__ == "__main__":
    test_torch()

I obtain these errors:

TypeError: <class 'keras.src.layers.convolutional.conv2d.Conv2D'> 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.layers', 'class_name': 'Conv2D', 'config': {'name': 'conv2d_20', 'trainable': True, 'dtype': 'float32', 'batch_input_shape': [None, 50, 50, 1], 'filters': 32, 'kernel_size': [3, 3], 'strides': [1, 1], 'padding': 'valid', 'data_format': 'channels_last', 'dilation_rate': [1, 1], 'groups': 1, 'activation': 'relu', 'use_bias': True, 'kernel_initializer': {'module': 'keras.initializers', 'class_name': 'VarianceScaling', 'config': {'scale': 1.0, 'mode': 'fan_avg', 'distribution': 'uniform', 'seed': None}, 'registered_name': None}, 'bias_initializer': {'module': 'keras.initializers', 'class_name': 'Zeros', 'config': {}, 'registered_name': None}, 'kernel_regularizer': None, 'bias_regularizer': None, 'activity_regularizer': None, 'kernel_constraint': None, 'bias_constraint': None}, 'registered_name': None, 'build_config': {'input_shape': [None, 50, 50, 1]}}.

Exception encountered: Error when deserializing class 'Conv2D' using config={'name': 'conv2d_20', 'trainable': True, 'dtype': 'float32', 'batch_input_shape': [None, 50, 50, 1], 'filters': 32, 'kernel_size': [3, 3], 'strides': [1, 1], 'padding': 'valid', 'data_format': 'channels_last', 'dilation_rate': [1, 1], 'groups': 1, 'activation': 'relu', 'use_bias': True, 'kernel_initializer': {'module': 'keras.initializers', 'class_name': 'VarianceScaling', 'config': {'scale': 1.0, 'mode': 'fan_avg', 'distribution': 'uniform', 'seed': None}, 'registered_name': None}, 'bias_initializer': {'module': 'keras.initializers', 'class_name': 'Zeros', 'config': {}, 'registered_name': None}, 'kernel_regularizer': None, 'bias_regularizer': None, 'activity_regularizer': None, 'kernel_constraint': None, 'bias_constraint': None}.

Exception encountered: Unrecognized keyword arguments passed to Conv2D: {'batch_input_shape': [None, 50, 50, 1]}

and

TypeError: <class 'keras.src.models.sequential.Sequential'> 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', 'class_name': 'Sequential', 'config': {'name': 'sequential', 'layers': [{'module': 'keras.layers', 'class_name': 'InputLayer', 'config': {'batch_input_shape': [None, 50, 50, 1], 'dtype': 'float32', 'sparse': False, 'ragged': False, 'name': 'conv2d_20_input'}, 'registered_name': None}, {'module': 'keras.layers', 'class_name': 'Conv2D', 'config': {'name': 'conv2d_20', 'trainable': True, 'dtype': 'float32', 'batch_input_shape': [None, 50, 50, 1], 'filters': 32, 'kernel_size': [3, 3], 'strides': [1, 1], 'padding': 'valid', 'data_format': 'channels_last', 'dilation_rate': [1, 1], 'groups': 1, 'activation': 'relu', 'use_bias': True, 'kernel_initializer': {'module': 'keras.initializers', 'class_name': 'VarianceScaling', 'config': {'scale': 1.0, 'mode': 'fan_avg', 'distribution': 'uniform', 'seed': None}, 'registered_name': None}, 'bias_initializer': {'module': 'keras.initializers', 'class_name': 'Zeros', 'config': {}, 'registered_name': None}, 'kernel_regularizer': None, 'bias_regularizer': None, 'activity_regularizer': None, 'kernel_constraint': None, 'bias_constraint': None}, 'registered_name': None, 'build_config': {'input_shape': [None, 50, 50, 1]}}, {'module': 'keras.layers', 'class_name': 'Activation', 'config': {'name': 'activation_13', 'trainable': True, 'dtype': 'float32', 'activation': 'relu'}, 'registered_name': None, 'build_config': {'input_shape': [None, 48, 48, 32]}}, {'module': 'keras.layers', 'class_name': 'Conv2D', 'config': {'name': 'conv2d_21', 'trainable': True, 'dtype': 'float32', 'filters': 32, 'kernel_size': [3, 3], 'strides': [1, 1], 'padding': 'valid', 'data_format': 'channels_last', 'dilation_rate': [1, 1], 'groups': 1, 'activation': 'linear', 'use_bias': True, 'kernel_initializer': {'module': 'keras.initializers', 'class_name': 'VarianceScaling', 'config': {'scale': 1.0, 'mode': 'fan_avg', 'distribution': 'uniform', 'seed': None}, 'registered_name': None}, 'bias_initializer': {'module': 'keras.initializers', 'class_name': 'Zeros', 'config': {}, 'registered_name': None}, 'kernel_regularizer': None, 'bias_regularizer': None, 'activity_regularizer': None, 'kernel_constraint': None, 'bias_constraint': None}, 'registered_name': None, 'build_config': {'input_shape': [None, 48, 48, 32]}}, {'module': 'keras.layers', 'class_name': 'Activation', 'config': {'name': 'activation_14', 'trainable': True, 'dtype': 'float32', 'activation': 'relu'}, 'registered_name': None, 'build_config': {'input_shape': [None, 46, 46, 32]}}, {'module': 'keras.layers', 'class_name': 'MaxPooling2D', 'config': {'name': 'max_pooling2d_10', 'trainable': True, 'dtype': 'float32', 'pool_size': [2, 2], 'padding': 'valid', 'strides': [2, 2], 'data_format': 'channels_last'}, 'registered_name': None, 'build_config': {'input_shape': [None, 46, 46, 32]}}, {'module': 'keras.layers', 'class_name': 'Dropout', 'config': {'name': 'dropout_17', 'trainable': True, 'dtype': 'float32', 'rate': 0.25, 'noise_shape': None, 'seed': None}, 'registered_name': None, 'build_config': {'input_shape': [None, 23, 23, 32]}}, {'module': 'keras.layers', 'class_name': 'Conv2D', 'config': {'name': 'conv2d_22', 'trainable': True, 'dtype': 'float32', 'filters': 64, 'kernel_size': [3, 3], 'strides': [1, 1], 'padding': 'same', 'data_format': 'channels_last', 'dilation_rate': [1, 1], 'groups': 1, 'activation': 'linear', 'use_bias': True, 'kernel_initializer': {'module': 'keras.initializers', 'class_name': 'VarianceScaling', 'config': {'scale': 1.0, 'mode': 'fan_avg', 'distribution': 'uniform', 'seed': None}, 'registered_name': None}, 'bias_initializer': {'module': 'keras.initializers', 'class_name': 'Zeros', 'config': {}, 'registered_name': None}, 'kernel_regularizer': None, 'bias_regularizer': None, 'activity_regularizer': None, 'kernel_constraint': None, 'bias_constraint': None}, 'registered_name': None, 'build_config': {'input_shape': [None, 23, 23, 32]}}, {'module': 'keras.layers', 'class_name': 'Activation', 'config': {'name': 'activation_15', 'trainable': True, 'dtype': 'float32', 'activation': 'relu'}, 'registered_name': None, 'build_config': {'input_shape': [None, 23, 23, 64]}}, {'module': 'keras.layers', 'class_name': 'Conv2D', 'config': {'name': 'conv2d_23', 'trainable': True, 'dtype': 'float32', 'filters': 64, 'kernel_size': [3, 3], 'strides': [1, 1], 'padding': 'valid', 'data_format': 'channels_last', 'dilation_rate': [1, 1], 'groups': 1, 'activation': 'linear', 'use_bias': True, 'kernel_initializer': {'module': 'keras.initializers', 'class_name': 'VarianceScaling', 'config': {'scale': 1.0, 'mode': 'fan_avg', 'distribution': 'uniform', 'seed': None}, 'registered_name': None}, 'bias_initializer': {'module': 'keras.initializers', 'class_name': 'Zeros', 'config': {}, 'registered_name': None}, 'kernel_regularizer': None, 'bias_regularizer': None, 'activity_regularizer': None, 'kernel_constraint': None, 'bias_constraint': None}, 'registered_name': None, 'build_config': {'input_shape': [None, 23, 23, 64]}}, {'module': 'keras.layers', 'class_name': 'Activation', 'config': {'name': 'activation_16', 'trainable': True, 'dtype': 'float32', 'activation': 'relu'}, 'registered_name': None, 'build_config': {'input_shape': [None, 21, 21, 64]}}, {'module': 'keras.layers', 'class_name': 'MaxPooling2D', 'config': {'name': 'max_pooling2d_11', 'trainable': True, 'dtype': 'float32', 'pool_size': [2, 2], 'padding': 'valid', 'strides': [2, 2], 'data_format': 'channels_last'}, 'registered_name': None, 'build_config': {'input_shape': [None, 21, 21, 64]}}, {'module': 'keras.layers', 'class_name': 'Dropout', 'config': {'name': 'dropout_18', 'trainable': True, 'dtype': 'float32', 'rate': 0.25, 'noise_shape': None, 'seed': None}, 'registered_name': None, 'build_config': {'input_shape': [None, 10, 10, 64]}}, {'module': 'keras.layers', 'class_name': 'Flatten', 'config': {'name': 'flatten_8', 'trainable': True, 'dtype': 'float32', 'data_format': 'channels_last'}, 'registered_name': None, 'build_config': {'input_shape': [None, 10, 10, 64]}}, {'module': 'keras.layers', 'class_name': 'Dense', 'config': {'name': 'dense_15', 'trainable': True, 'dtype': 'float32', 'units': 512, 'activation': 'linear', 'use_bias': True, 'kernel_initializer': {'module': 'keras.initializers', 'class_name': 'VarianceScaling', 'config': {'scale': 1.0, 'mode': 'fan_avg', 'distribution': 'uniform', 'seed': None}, 'registered_name': None}, 'bias_initializer': {'module': 'keras.initializers', 'class_name': 'Zeros', 'config': {}, 'registered_name': None}, 'kernel_regularizer': None, 'bias_regularizer': None, 'activity_regularizer': None, 'kernel_constraint': None, 'bias_constraint': None}, 'registered_name': None, 'build_config': {'input_shape': [None, 6400]}}, {'module': 'keras.layers', 'class_name': 'Activation', 'config': {'name': 'activation_17', 'trainable': True, 'dtype': 'float32', 'activation': 'relu'}, 'registered_name': None, 'build_config': {'input_shape': [None, 512]}}, {'module': 'keras.layers', 'class_name': 'Dropout', 'config': {'name': 'dropout_19', 'trainable': True, 'dtype': 'float32', 'rate': 0.5, 'noise_shape': None, 'seed': None}, 'registered_name': None, 'build_config': {'input_shape': [None, 512]}}, {'module': 'keras.layers', 'class_name': 'Dense', 'config': {'name': 'dense_16', 'trainable': True, 'dtype': 'float32', 'units': 2, 'activation': 'linear', 'use_bias': True, 'kernel_initializer': {'module': 'keras.initializers', 'class_name': 'VarianceScaling', 'config': {'scale': 1.0, 'mode': 'fan_avg', 'distribution': 'uniform', 'seed': None}, 'registered_name': None}, 'bias_initializer': {'module': 'keras.initializers', 'class_name': 'Zeros', 'config': {}, 'registered_name': None}, 'kernel_regularizer': None, 'bias_regularizer': None, 'activity_regularizer': None, 'kernel_constraint': None, 'bias_constraint': None}, 'registered_name': None, 'build_config': {'input_shape': [None, 512]}}, {'module': 'keras.layers', 'class_name': 'Activation', 'config': {'name': 'activation_18', 'trainable': True, 'dtype': 'float32', 'activation': 'softmax'}, 'registered_name': None, 'build_config': {'input_shape': [None, 2]}}]}, 'registered_name': None, 'build_config': {'input_shape': [None, 50, 50, 1]}, 'compile_config': None}.

Exception encountered: <class 'keras.src.layers.convolutional.conv2d.Conv2D'> 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.layers', 'class_name': 'Conv2D', 'config': {'name': 'conv2d_20', 'trainable': True, 'dtype': 'float32', 'batch_input_shape': [None, 50, 50, 1], 'filters': 32, 'kernel_size': [3, 3], 'strides': [1, 1], 'padding': 'valid', 'data_format': 'channels_last', 'dilation_rate': [1, 1], 'groups': 1, 'activation': 'relu', 'use_bias': True, 'kernel_initializer': {'module': 'keras.initializers', 'class_name': 'VarianceScaling', 'config': {'scale': 1.0, 'mode': 'fan_avg', 'distribution': 'uniform', 'seed': None}, 'registered_name': None}, 'bias_initializer': {'module': 'keras.initializers', 'class_name': 'Zeros', 'config': {}, 'registered_name': None}, 'kernel_regularizer': None, 'bias_regularizer': None, 'activity_regularizer': None, 'kernel_constraint': None, 'bias_constraint': None}, 'registered_name': None, 'build_config': {'input_shape': [None, 50, 50, 1]}}.

Exception encountered: Error when deserializing class 'Conv2D' using config={'name': 'conv2d_20', 'trainable': True, 'dtype': 'float32', 'batch_input_shape': [None, 50, 50, 1], 'filters': 32, 'kernel_size': [3, 3], 'strides': [1, 1], 'padding': 'valid', 'data_format': 'channels_last', 'dilation_rate': [1, 1], 'groups': 1, 'activation': 'relu', 'use_bias': True, 'kernel_initializer': {'module': 'keras.initializers', 'class_name': 'VarianceScaling', 'config': {'scale': 1.0, 'mode': 'fan_avg', 'distribution': 'uniform', 'seed': None}, 'registered_name': None}, 'bias_initializer': {'module': 'keras.initializers', 'class_name': 'Zeros', 'config': {}, 'registered_name': None}, 'kernel_regularizer': None, 'bias_regularizer': None, 'activity_regularizer': None, 'kernel_constraint': None, 'bias_constraint': None}.

Exception encountered: Unrecognized keyword arguments passed to Conv2D: {'batch_input_shape': [None, 50, 50, 1]}

With this, I am wondering

1. Is there a way to access this model without retraining it?
2. For my first code file/portion, should I save via keras before I compile the model or before I put the weights (aka when should I save the model in these steps)?
3. Worst case, can I load the model from the json file, do model.summary(), then recreate the model in Keras 3 than add the weights from the .h5 file that was made in tensorflow 2? Will that be fine?

This was a similar issue to this StackOverflow question here: Similar Issue

[newresu]

• Try: 1. Load with tf 2.15 (installs Keras 2 by default, but check the version.), 2. save it as .keras 3. re-load it with Keras 3?

.keras was there in 2.15.

---

From the error log the serialisation just changed this key name:

Exception encountered: Unrecognized keyword arguments passed to Conv2D: {'batch_input_shape': [None, 50, 50, 1]}

[SamanehSaadat]

Hi @mannypaeza!

As @ghsanti recommended, you can load your model in Keras 2, save it in .keras format and then load it in Keras 3.

[github-actions[bot]]

This issue is stale because it has been open for 14 days with no activity. It will be closed if no further activity occurs. Thank you.

[github-actions[bot]]

This issue was closed because it has been inactive for 28 days. Please reopen if you'd like to work on this further.

[google-ml-butler[bot]]

Are you satisfied with the resolution of your issue? Yes No