feat: bump tensorflow and tensorflow probability versions (#336)

* feat: bump tensorflow and tensorflow probability versions

* chore: update github flows python version

* chore: update github flows python version for pull request

Author: fabclmnt

.github/workflows/prerelease.yml

@@ -19,10 +19,10 @@ jobs:
       id: version
       run: echo ::set-output name=value::${GITHUB_REF#refs/*/}
 
-    - name: Setup Python 3.8
+    - name: Setup Python 3.10
       uses: actions/setup-python@v5
       with:
-        python-version: '3.8'
+        python-version: '3.10'
 
     - name: Install dependencies
       run: |

.github/workflows/pull_request.yml

@@ -17,10 +17,10 @@ jobs:
     steps:
     - uses: actions/checkout@v4
 
-    - name: Setup Python 3.8
+    - name: Setup Python 3.10
       uses: actions/setup-python@v5
       with:
-        python-version: '3.8'
+        python-version: '3.10'
 
     - name: Cache pip
       id: cache

requirements.txt

@@ -1,12 +1,12 @@
-requests>=2.30, <2.31
+requests>=2.28, <2.31
 pandas<3
 numpy<2
 scikit-learn<2
 matplotlib<4
-tensorflow==2.12.0
-tensorflow-probability==0.19.0
+tensorflow==2.15.*
+tensorflow-probability[tf]
 easydict==1.10
 pmlb==1.0.*
 tqdm<5.0
-typeguard==4.0.*
+typeguard==4.2.*
 pytest==7.4.*

src/ydata_synthetic/synthesizers/saving_keras.py

@@ -1,19 +1,21 @@
+import tensorflow.python.keras as tf_keras
+from keras import __version__
+tf_keras.__version__ = __version__
+
 from tensorflow.keras import Model
-from tensorflow.python.keras.layers import deserialize, serialize
-from tensorflow.python.keras.saving import saving_utils
 
 def unpack(model, training_config, weights):
-    restored_model = deserialize(model)
+    restored_model = tf_keras.layers.deserialize(model)
     if training_config is not None:
-        restored_model.compile(**saving_utils.compile_args_from_training_config(training_config))
+        restored_model.compile(**tf_keras.saving.saving_utils.compile_args_from_training_config(training_config))
     restored_model.set_weights(weights)
     return restored_model
 
 def make_keras_picklable():
     def __reduce__(self):
-        model_metadata = saving_utils.model_metadata(self)
+        model_metadata = tf_keras.saving.saving_utils.model_metadata(self)
         training_config = model_metadata.get("training_config", None)
-        model = serialize(self)
+        model = tf_keras.layers.serialize(self)
         weights = self.get_weights()
         return (unpack, (model, training_config, weights))