TypeError: Error when deserializing class 'Embedding' using config

[jmalhot]

All of sudden today, I started getting following error while using ethnicolr.pred_fl_reg_ln (version 0.9.6)

Error Category: UNCLASSIFIED_ERROR; TypeError: Error when deserializing class 'Embedding' using config={'name': 'embedding', 'trainable': True, 'dtype': 'float32', 'input_dim': 1152, 'output_dim': 32, 'embeddings_initializer': {'class_name': 'RandomUniform', 'config': {'minval': -0.05, 'maxval': 0.05, 'seed': None}}, 'embeddings_regularizer': None, 'activity_regularizer': None, 'embeddings_constraint': None, 'mask_zero': False, 'input_length': 20, 'input_shape': [None, 20]}.

I was able to work with no issues last week. Has anything changed in the source code recently ?

[soodoku]

No change to source code. I am assuming it is some versioning issue at your end. Try giving me a reproducible example so that we can test at our end

…

On Mon, Mar 11, 2024, 1:54 PM Jatin Malhotra ***@***.***> wrote: All of sudden today, I started getting following error while using ethnicolr.pred_fl_reg_ln (version 0.9.6) Error Category: UNCLASSIFIED_ERROR; TypeError: Error when deserializing class 'Embedding' using config={'name': 'embedding', 'trainable': True, 'dtype': 'float32', 'input_dim': 1152, 'output_dim': 32, 'embeddings_initializer': {'class_name': 'RandomUniform', 'config': {'minval': -0.05, 'maxval': 0.05, 'seed': None}}, 'embeddings_regularizer': None, 'activity_regularizer': None, 'embeddings_constraint': None, 'mask_zero': False, 'input_length': 20, 'input_shape': [None, 20]}. I was able to work with no issues last week. Has anything changed in the source code recently ? — Reply to this email directly, view it on GitHub <#93>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAFQEOUBU5IVYHSCK2MPQZ3YXYRYRAVCNFSM6AAAAABERA7YTKVHI2DSMVQWIX3LMV43ASLTON2WKOZSGE4DAMRRGE2DQOA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[jmalhot]

Thanks for your quick response. Following are the steps to reproduce the issue:-

##Create a new python vitual env
python -m venv ethnicolr_testing

##Activate virtual envrionment
source ethnicolr_testing/bin/activate

##Install ethnicolr
pip install ethnicolr==0.9.6

##Run the python script

import pandas as pd
from ethnicolr import (pred_fl_reg_ln)
 data = {
    'Name': ['Yan Lee', 
             'Mohammad Khan'
             ]
}
df = pd.DataFrame(data)
odf = pred_fl_reg_ln(df, 'Name', num_iter=100, conf_int=0.9)
print(odf)

##Following is the error returned

Traceback (most recent call last):
  File "/Users/user/Downloads/ethnicolr_testing/lib/python3.9/site-packages/keras/src/ops/operation.py", line 196, in from_config
    return cls(**config)
  File "/Users/user/Downloads/ethnicolr_testing/lib/python3.9/site-packages/keras/src/layers/core/embedding.py", line 81, in __init__
    super().__init__(**kwargs)
  File "/Users/user/Downloads/ethnicolr_testing/lib/python3.9/site-packages/keras/src/layers/layer.py", line 265, in __init__
    raise ValueError(
ValueError: Unrecognized keyword arguments passed to Embedding: {'input_length': 20}

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/Users/user/Downloads/test.py", line 18, in <module>
    odf = pred_fl_reg_ln(df, 'Name', num_iter=100, conf_int=0.9)
  File "/Users/user/Downloads/ethnicolr_testing/lib/python3.9/site-packages/ethnicolr/pred_fl_reg_ln.py", line 42, in pred_fl_reg_ln
    rdf = cls.transform_and_pred(df=df,
  File "/Users/user/Downloads/ethnicolr_testing/lib/python3.9/site-packages/ethnicolr/ethnicolr_class.py", line 108, in transform_and_pred
    cls.model = load_model(MODEL)
  File "/Users/user/Downloads/ethnicolr_testing/lib/python3.9/site-packages/keras/src/saving/saving_api.py", line 183, in load_model
    return legacy_h5_format.load_model_from_hdf5(filepath)
  File "/Users/user/Downloads/ethnicolr_testing/lib/python3.9/site-packages/keras/src/legacy/saving/legacy_h5_format.py", line 133, in load_model_from_hdf5
    model = saving_utils.model_from_config(
  File "/Users/user/Downloads/ethnicolr_testing/lib/python3.9/site-packages/keras/src/legacy/saving/saving_utils.py", line 85, in model_from_config
    return serialization.deserialize_keras_object(
  File "/Users/user/Downloads/ethnicolr_testing/lib/python3.9/site-packages/keras/src/legacy/saving/serialization.py", line 495, in deserialize_keras_object
    deserialized_obj = cls.from_config(
  File "/Users/user/Downloads/ethnicolr_testing/lib/python3.9/site-packages/keras/src/models/sequential.py", line 327, in from_config
    layer = saving_utils.model_from_config(
  File "/Users/user/Downloads/ethnicolr_testing/lib/python3.9/site-packages/keras/src/legacy/saving/saving_utils.py", line 85, in model_from_config
    return serialization.deserialize_keras_object(
  File "/Users/user/Downloads/ethnicolr_testing/lib/python3.9/site-packages/keras/src/legacy/saving/serialization.py", line 504, in deserialize_keras_object
    deserialized_obj = cls.from_config(cls_config)
  File "/Users/user/Downloads/ethnicolr_testing/lib/python3.9/site-packages/keras/src/ops/operation.py", line 198, in from_config
    raise TypeError(
TypeError: Error when deserializing class 'Embedding' using config={'name': 'embedding', 'trainable': True, 'dtype': 'float32', 'input_dim': 1152, 'output_dim': 32, 'embeddings_initializer': {'class_name': 'RandomUniform', 'config': {'minval': -0.05, 'maxval': 0.05, 'seed': None}}, 'embeddings_regularizer': None, 'activity_regularizer': None, 'embeddings_constraint': None, 'mask_zero': False, 'input_length': 20, 'input_shape': [None, 20]}.

Exception encountered: Unrecognized keyword arguments passed to Embedding: {'input_length': 20}
(ethnicolr_testing) bash-3.2$ 

[jmalhot]

Issue is Fixed:
pip install ethnicolr==0.9.6 by default was installing tensorflow 2.16.x which is incompatible with ethnicolr==0.9.6. Comptaible version is tensorflow version is 2.15.0

Please update the documentation