renamed

src/transformers/modeling_generation.py

@@ -25,9 +25,9 @@ from torch.nn import functional as F
 logger = logging.getLogger(__name__)
 
 
-class ModuleGenerationMixin:
+class GenerationMixin:
     """
-    A class contraining all of the functions supporting generation, to be used as a mixin.
+    A class contraining all of the functions supporting generation, to be used as a mixin in PreTrainedModel.
     """
 
     def prepare_inputs_for_generation(self, input_ids, **kwargs):

src/transformers/modeling_tf_generation.py

@@ -23,9 +23,9 @@ import tensorflow as tf
 logger = logging.getLogger(__name__)
 
 
-class TFModelGenerationMixin:
+class TFGenerationMixin:
     """
-    A class contraining all of the functions supporting generation, to be used as a mixin.
+    A class contraining all of the functions supporting generation, to be used as a mixin in TFPreTrainedModel.
     """
 
     def prepare_inputs_for_generation(self, inputs, **kwargs):

src/transformers/modeling_tf_utils.py

@@ -25,7 +25,7 @@ from tensorflow.python.keras.saving import hdf5_format
 
 from .configuration_utils import PretrainedConfig
 from .file_utils import DUMMY_INPUTS, TF2_WEIGHTS_NAME, WEIGHTS_NAME, cached_path, hf_bucket_url, is_remote_url
-from .modeling_tf_generation import TFModelGenerationMixin, shape_list
+from .modeling_tf_generation import TFGenerationMixin, shape_list
 from .modeling_tf_pytorch_utils import load_pytorch_checkpoint_in_tf2_model
 
 
@@ -145,7 +145,7 @@ class TFSequenceClassificationLoss:
 TFMultipleChoiceLoss = TFSequenceClassificationLoss
 
 
-class TFPreTrainedModel(tf.keras.Model, TFModelUtilsMixin, TFModelGenerationMixin):
+class TFPreTrainedModel(tf.keras.Model, TFModelUtilsMixin, TFGenerationMixin):
     r""" Base class for all TF models.
 
         :class:`~transformers.TFPreTrainedModel` takes care of storing the configuration of the models and handles methods for loading/downloading/saving models

src/transformers/modeling_utils.py

@@ -35,7 +35,7 @@ from .file_utils import (
     hf_bucket_url,
     is_remote_url,
 )
-from .modeling_generation import ModuleGenerationMixin
+from .modeling_generation import GenerationMixin
 
 
 logger = logging.getLogger(__name__)
@@ -262,7 +262,7 @@ class ModuleUtilsMixin:
         return head_mask
 
 
-class PreTrainedModel(nn.Module, ModuleUtilsMixin, ModuleGenerationMixin):
+class PreTrainedModel(nn.Module, ModuleUtilsMixin, GenerationMixin):
     r""" Base class for all models.
 
         :class:`~transformers.PreTrainedModel` takes care of storing the configuration of the models and handles methods for loading/downloading/saving models