Instantiate model only once in pipeline (#10888)

* Instantiate model only once in pipeline

* Remove documentation of deprecated method

* Add FutureWarning

* Update src/transformers/pipelines/base.py

Co-authored-by: Lysandre Debut <lysandre@huggingface.co>

Co-authored-by: Lysandre Debut <lysandre@huggingface.co>

docs/source/internal/pipelines_utils.rst

@@ -47,6 +47,4 @@ Data format
 Utilities
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. autofunction:: transformers.pipelines.get_framework
-
 .. autoclass:: transformers.pipelines.PipelineException

src/transformers/pipelines/__init__.py

@@ -34,7 +34,7 @@ from .base import (
     PipelineDataFormat,
     PipelineException,
     get_default_model,
-    get_framework,
+    infer_framework_from_model,
 )
 from .conversational import Conversation, ConversationalPipeline
 from .feature_extraction import FeatureExtractionPipeline
@@ -341,10 +341,6 @@ def pipeline(
         # At that point framework might still be undetermined
         model = get_default_model(targeted_task, framework, task_options)
 
-    framework = framework or get_framework(model)
-
-    task_class, model_class = targeted_task["impl"], targeted_task[framework]
-
     # Try to infer tokenizer from model or config name (if provided as str)
     if tokenizer is None:
         if isinstance(model, str):
@@ -365,6 +361,12 @@ def pipeline(
     elif isinstance(config, str):
         modelcard = config
 
+    # Infer the framework form the model
+    if framework is None:
+        framework, model = infer_framework_from_model(model, targeted_task, revision=revision)
+
+    task_class, model_class = targeted_task["impl"], targeted_task[framework]
+
     # Instantiate tokenizer if needed
     if isinstance(tokenizer, (str, tuple)):
         if isinstance(tokenizer, tuple):
@@ -406,14 +408,13 @@ def pipeline(
             )
 
         model = model_class.from_pretrained(model, config=config, revision=revision, **model_kwargs)
+
     if task == "translation" and model.config.task_specific_params:
         for key in model.config.task_specific_params:
             if key.startswith("translation"):
                 task = key
                 warnings.warn(
-                        '"translation" task was used, instead of "translation_XX_to_YY", defaulting to "{}"'.format(
+                    f'"translation" task was used, instead of "translation_XX_to_YY", defaulting to "{task}"',
-                            task
-                        ),
                     UserWarning,
                 )
                 break

src/transformers/pipelines/base.py

@@ -17,6 +17,7 @@ import json
 import os
 import pickle
 import sys
+import warnings
 from abc import ABC, abstractmethod
 from contextlib import contextmanager
 from os.path import abspath, exists
@@ -46,6 +47,55 @@ if TYPE_CHECKING:
 logger = logging.get_logger(__name__)
 
 
+def infer_framework_from_model(model, model_classes: Optional[Dict[str, type]] = None, revision: Optional[str] = None):
+    """
+    Select framework (TensorFlow or PyTorch) to use from the :obj:`model` passed. Returns a tuple (framework, model).
+
+    If :obj:`model` is instantiated, this function will just infer the framework from the model class. Otherwise
+    :obj:`model` is actually a checkpoint name and this method will try to instantiate it using :obj:`model_classes`.
+    Since we don't want to instantiate the model twice, this model is returned for use by the pipeline.
+
+    If both frameworks are installed and available for :obj:`model`, PyTorch is selected.
+
+    Args:
+        model (:obj:`str`, :class:`~transformers.PreTrainedModel` or :class:`~transformers.TFPreTrainedModel`):
+            The model to infer the framework from. If :obj:`str`, a checkpoint name. The model to infer the framewrok
+            from.
+        model_classes (dictionary :obj:`str` to :obj:`type`, `optional`):
+            A mapping framework to class.
+        revision (:obj:`str`, `optional`):
+            The specific model version to use. It can be a branch name, a tag name, or a commit id, since we use a
+            git-based system for storing models and other artifacts on huggingface.co, so ``revision`` can be any
+            identifier allowed by git.
+
+    Returns:
+        :obj:`Tuple`: A tuple framework, model.
+    """
+    if not is_tf_available() and not is_torch_available():
+        raise RuntimeError(
+            "At least one of TensorFlow 2.0 or PyTorch should be installed. "
+            "To install TensorFlow 2.0, read the instructions at https://www.tensorflow.org/install/ "
+            "To install PyTorch, read the instructions at https://pytorch.org/."
+        )
+    if isinstance(model, str):
+        if is_torch_available() and not is_tf_available():
+            model_class = model_classes.get("pt", AutoModel)
+            model = model_class.from_pretrained(model, revision=revision)
+        elif is_tf_available() and not is_torch_available():
+            model_class = model_classes.get("tf", TFAutoModel)
+            model = model_class.from_pretrained(model, revision=revision)
+        else:
+            try:
+                model_class = model_classes.get("pt", AutoModel)
+                model = model_class.from_pretrained(model, revision=revision)
+            except OSError:
+                model_class = model_classes.get("tf", TFAutoModel)
+                model = model_class.from_pretrained(model, revision=revision)
+
+    framework = "tf" if model.__class__.__name__.startswith("TF") else "pt"
+    return framework, model
+
+
 def get_framework(model, revision: Optional[str] = None):
     """
     Select framework (TensorFlow or PyTorch) to use.
@@ -55,6 +105,10 @@ def get_framework(model, revision: Optional[str] = None):
             If both frameworks are installed, picks the one corresponding to the model passed (either a model class or
             the model name). If no specific model is provided, defaults to using PyTorch.
     """
+    warnings.warn(
+        "`get_framework` is deprecated and will be removed in v5, use `infer_framework_from_model` instead.",
+        FutureWarning,
+    )
     if not is_tf_available() and not is_torch_available():
         raise RuntimeError(
             "At least one of TensorFlow 2.0 or PyTorch should be installed. "
@@ -474,7 +528,7 @@ class Pipeline(_ScikitCompat):
     ):
 
         if framework is None:
-            framework = get_framework(model)
+            framework = infer_framework_from_model(model)
 
         self.task = task
         self.model = model