Improved keras imports (#24448)

* An end to accursed version-specific imports

* No more K.is_keras_tensor() either

* Update dependency tables

* Use a cleaner call context function getter

* Add a cap to <2.14

* Add cap to examples requirements too

examples/tensorflow/_tests_requirements.txt

-tensorflow<2.13
+tensorflow<2.14
 tensorboard
 scikit-learn
 seqeval

setup.py

@@ -168,9 +168,9 @@ _deps = [
     "sudachipy>=0.6.6",
     "sudachidict_core>=20220729",
     # TensorFlow pin. When changing this value, update examples/tensorflow/_tests_requirements.txt accordingly
-    "tensorflow-cpu>=2.4,<2.13",
-    "tensorflow>=2.4,<2.13",
-    "tensorflow-text<2.13",
+    "tensorflow-cpu>=2.6,<2.14",
+    "tensorflow>=2.6,<2.14",
+    "tensorflow-text<2.14",
     "tf2onnx",
     "timeout-decorator",
     "timm",

src/transformers/dependency_versions_table.py

@@ -72,9 +72,9 @@ deps = {
     "starlette": "starlette",
     "sudachipy": "sudachipy>=0.6.6",
     "sudachidict_core": "sudachidict_core>=20220729",
-    "tensorflow-cpu": "tensorflow-cpu>=2.4,<2.13",
-    "tensorflow": "tensorflow>=2.4,<2.13",
-    "tensorflow-text": "tensorflow-text<2.13",
+    "tensorflow-cpu": "tensorflow-cpu>=2.6,<2.14",
+    "tensorflow": "tensorflow>=2.6,<2.14",
+    "tensorflow-text": "tensorflow-text<2.14",
     "tf2onnx": "tf2onnx",
     "timeout-decorator": "timeout-decorator",
     "timm": "timm",

src/transformers/modeling_tf_pytorch_utils.py

@@ -251,12 +251,7 @@ def load_pytorch_state_dict_in_tf2_model(
     """Load a pytorch state_dict in a TF 2.0 model. pt_state_dict can be either an actual dict or a lazy-loading
     safetensors archive created with the safe_open() function."""
     import tensorflow as tf
-    from packaging.version import parse
-
-    if parse(tf.__version__) >= parse("2.11.0"):
-        from keras import backend as K
-    else:
-        from tensorflow.python.keras import backend as K
+    from keras import backend as K
 
     if tf_inputs is None:
         tf_inputs = tf_model.dummy_inputs

src/transformers/modeling_tf_utils.py

@@ -33,7 +33,9 @@ import h5py
 import numpy as np
 import tensorflow as tf
 from huggingface_hub import Repository, list_repo_files
+from keras import backend as K
 from packaging.version import parse
+from tensorflow.python.util.keras_deps import get_call_context_function
 
 from . import DataCollatorWithPadding, DefaultDataCollator
 from .activations_tf import get_tf_activation
@@ -71,20 +73,6 @@ from .utils import (
 from .utils.hub import convert_file_size_to_int, get_checkpoint_shard_files
 
 
-if parse(tf.__version__).minor >= 13:
-    from keras import backend as K
-    from keras.__internal__ import KerasTensor
-    from keras.src.engine.base_layer_utils import call_context
-elif parse(tf.__version__).minor >= 11:
-    from keras import backend as K
-    from keras.engine.base_layer_utils import call_context
-    from keras.engine.keras_tensor import KerasTensor
-else:
-    from tensorflow.python.keras import backend as K
-    from tensorflow.python.keras.engine.base_layer_utils import call_context
-    from tensorflow.python.keras.engine.keras_tensor import KerasTensor
-
-
 if is_safetensors_available():
     from safetensors import safe_open
     from safetensors.tensorflow import save_file as safe_save_file
@@ -99,13 +87,10 @@ tf_logger = tf.get_logger()
 TFModelInputType = Union[
     List[tf.Tensor],
     List[np.ndarray],
-    List[KerasTensor],
     Dict[str, tf.Tensor],
     Dict[str, np.ndarray],
-    Dict[str, KerasTensor],
     tf.Tensor,
     np.ndarray,
-    KerasTensor,
 ]
 
 
@@ -472,7 +457,7 @@ def input_processing(func, config, **kwargs):
     main_input_name = parameter_names[0]
     main_input = kwargs.pop(main_input_name, None)
     output = {}
-    allowed_types = (tf.Tensor, bool, int, ModelOutput, tuple, list, dict, np.ndarray, KerasTensor)
+    allowed_types = (tf.Tensor, bool, int, ModelOutput, tuple, list, dict, np.ndarray)
 
     if "inputs" in kwargs["kwargs_call"]:
         warnings.warn(
@@ -511,7 +496,7 @@ def input_processing(func, config, **kwargs):
         kwargs.pop("kwargs_call")
 
     for k, v in kwargs.items():
-        if isinstance(v, allowed_types) or v is None:
+        if isinstance(v, allowed_types) or tf.is_tensor(v) or v is None:
             output[k] = v
         else:
             raise ValueError(f"Data of type {type(v)} is not allowed only {allowed_types} is accepted for {k}.")
@@ -564,7 +549,7 @@ def input_processing(func, config, **kwargs):
             else:
                 raise ValueError(f"Data of type {type(v)} is not allowed only {allowed_types} is accepted for {k}.")
     else:
-        if isinstance(main_input, (tf.Tensor, KerasTensor)) or main_input is None:
+        if tf.is_tensor(main_input) or main_input is None:
             output[main_input_name] = main_input
         else:
             raise ValueError(
@@ -1142,6 +1127,7 @@ class TFPreTrainedModel(tf.keras.Model, TFModelUtilsMixin, TFGenerationMixin, Pu
         return "tf"
 
     def build(self, input_shape=None):
+        call_context = get_call_context_function()
         if self.built or call_context().in_call:
             self.built = True
         else: