Enable ONNX export for transformers 4.45 (#2045)

* Enable ONNX export for transformers 4.45

* add comment

* update setup

Author: echarlaix

optimum/exporters/onnx/convert.py

@@ -26,7 +26,6 @@
 
 import numpy as np
 import onnx
-import transformers
 from transformers.modeling_utils import get_parameter_dtype
 from transformers.utils import is_tf_available, is_torch_available
 
@@ -531,6 +530,11 @@ def export_pytorch(
     logger.info(f"Using framework PyTorch: {torch.__version__}")
     FORCE_ONNX_EXTERNAL_DATA = os.getenv("FORCE_ONNX_EXTERNAL_DATA", "0") == "1"
 
+    model_kwargs = model_kwargs or {}
+    # num_logits_to_keep was added in transformers 4.45 and isn't added as inputs when exporting the model
+    if check_if_transformers_greater("4.44.99") and "num_logits_to_keep" in signature(model.forward).parameters.keys():
+        model_kwargs["num_logits_to_keep"] = 0
+
     with torch.no_grad():
         model.config.return_dict = True
         model = model.eval()
@@ -1001,11 +1005,6 @@ def onnx_export_from_model(
     >>> onnx_export_from_model(model, output="gpt2_onnx/")
     ```
     """
-    if check_if_transformers_greater("4.44.99"):
-        raise ImportError(
-            f"ONNX conversion disabled for now for transformers version greater than v4.45, found {transformers.__version__}"
-        )
-
     TasksManager.standardize_model_attributes(model)
 
     if hasattr(model.config, "export_model_type"):

setup.py

@@ -54,7 +54,6 @@
         "datasets>=1.2.1",
         "evaluate",
         "protobuf>=3.20.1",
-        "transformers<4.45.0",
     ],
     "onnxruntime-gpu": [
         "onnx",
@@ -63,10 +62,9 @@
         "evaluate",
         "protobuf>=3.20.1",
         "accelerate",  # ORTTrainer requires it.
-        "transformers<4.45.0",
     ],
-    "exporters": ["onnx", "onnxruntime", "timm", "transformers<4.45.0"],
-    "exporters-gpu": ["onnx", "onnxruntime-gpu", "timm", "transformers<4.45.0"],
+    "exporters": ["onnx", "onnxruntime", "timm"],
+    "exporters-gpu": ["onnx", "onnxruntime-gpu", "timm"],
     "exporters-tf": [
         "tensorflow>=2.4,<=2.12.1",
         "tf2onnx",
@@ -77,7 +75,6 @@
         "numpy<1.24.0",
         "datasets<=2.16",
         "transformers[sentencepiece]>=4.26,<4.38",
-        "transformers<4.45.0",
     ],
     "diffusers": ["diffusers"],
     "intel": "optimum-intel>=1.18.0",