Add transformers v4.49 support (#2191)

* Add transformers v4.49 support * fix logits_to_keep_name * tmp use transformers version for tests * fix transformers incompatibility * revert
huggingface · Feb 24, 2025 · f4c9021 · f4c9021
1 parent ce533cf
commit f4c9021
Show file tree

Hide file tree

Showing 2 changed files with 9 additions and 7 deletions.
diff --git a/optimum/exporters/onnx/convert.py b/optimum/exporters/onnx/convert.py
@@ -512,8 +512,10 @@ def export_pytorch(
 
     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 is_transformers_version(">=", "4.44.99") and "num_logits_to_keep" in signature(model.forward).parameters.keys():
-        model_kwargs["num_logits_to_keep"] = 0
+    if is_transformers_version(">=", "4.45"):
+        logits_to_keep_name = "logits_to_keep" if is_transformers_version(">=", "4.49") else "num_logits_to_keep"
+        if logits_to_keep_name in signature(model.forward).parameters.keys():
+            model_kwargs[logits_to_keep_name] = 0
 
     with torch.no_grad():
         model.config.return_dict = True

diff --git a/setup.py b/setup.py
@@ -50,15 +50,15 @@
         "datasets>=1.2.1",
         "evaluate",
         "protobuf>=3.20.1",
-        "transformers>=4.36,<4.49.0",
+        "transformers>=4.36,<4.50.0",
     ],
     "onnxruntime-gpu": [
         "onnx",
         "onnxruntime-gpu>=1.11.0",
         "datasets>=1.2.1",
         "evaluate",
         "protobuf>=3.20.1",
-        "transformers>=4.36,<4.49.0",
+        "transformers>=4.36,<4.50.0",
     ],
     "onnxruntime-training": [
         "torch-ort",
@@ -67,19 +67,19 @@
         "accelerate",
         "evaluate",
         "protobuf>=3.20.1",
-        "transformers>=4.36,<4.49.0",
+        "transformers>=4.36,<4.50.0",
     ],
     "exporters": [
         "onnx",
         "onnxruntime",
         "timm",
-        "transformers>=4.36,<4.49.0",
+        "transformers>=4.36,<4.50.0",
     ],
     "exporters-gpu": [
         "onnx",
         "onnxruntime-gpu",
         "timm",
-        "transformers>=4.36,<4.49.0",
+        "transformers>=4.36,<4.50.0",
     ],
     "exporters-tf": [
         "tensorflow>=2.4,<=2.12.1",