Rename second input dimension from "sequence" to "num_channels" for CV models (#17976)

src/transformers/models/beit/configuration_beit.py

@@ -194,7 +194,7 @@ class BeitOnnxConfig(OnnxConfig):
     def inputs(self) -> Mapping[str, Mapping[int, str]]:
         return OrderedDict(
             [
-                ("pixel_values", {0: "batch", 1: "sequence"}),
+                ("pixel_values", {0: "batch", 1: "num_channels"}),
             ]
         )
 

src/transformers/models/convnext/configuration_convnext.py

@@ -117,7 +117,7 @@ class ConvNextOnnxConfig(OnnxConfig):
     def inputs(self) -> Mapping[str, Mapping[int, str]]:
         return OrderedDict(
             [
-                ("pixel_values", {0: "batch", 1: "sequence"}),
+                ("pixel_values", {0: "batch", 1: "num_channels"}),
             ]
         )
 

src/transformers/models/data2vec/configuration_data2vec_vision.py

@@ -193,7 +193,7 @@ class Data2VecVisionOnnxConfig(OnnxConfig):
     def inputs(self) -> Mapping[str, Mapping[int, str]]:
         return OrderedDict(
             [
-                ("pixel_values", {0: "batch", 1: "sequence"}),
+                ("pixel_values", {0: "batch", 1: "num_channels"}),
             ]
         )
 

src/transformers/models/deit/configuration_deit.py

@@ -137,7 +137,7 @@ class DeiTOnnxConfig(OnnxConfig):
     def inputs(self) -> Mapping[str, Mapping[int, str]]:
         return OrderedDict(
             [
-                ("pixel_values", {0: "batch", 1: "sequence"}),
+                ("pixel_values", {0: "batch", 1: "num_channels"}),
             ]
         )
 

src/transformers/models/detr/configuration_detr.py

@@ -220,8 +220,8 @@ class DetrOnnxConfig(OnnxConfig):
     def inputs(self) -> Mapping[str, Mapping[int, str]]:
         return OrderedDict(
             [
-                ("pixel_values", {0: "batch", 1: "sequence"}),
+                ("pixel_values", {0: "batch", 1: "num_channels"}),
-                ("pixel_mask", {0: "batch", 1: "sequence"}),
+                ("pixel_mask", {0: "batch"}),
             ]
         )
 

src/transformers/models/layoutlmv3/configuration_layoutlmv3.py

@@ -212,7 +212,7 @@ class LayoutLMv3OnnxConfig(OnnxConfig):
                     ("input_ids", {0: "batch", 1: "sequence"}),
                     ("bbox", {0: "batch", 1: "sequence"}),
                     ("attention_mask", {0: "batch", 1: "sequence"}),
-                    ("pixel_values", {0: "batch", 1: "sequence"}),
+                    ("pixel_values", {0: "batch", 1: "num_channels"}),
                 ]
             )
 

src/transformers/models/mobilevit/configuration_mobilevit.py

@@ -171,7 +171,7 @@ class MobileViTOnnxConfig(OnnxConfig):
 
     @property
     def inputs(self) -> Mapping[str, Mapping[int, str]]:
-        return OrderedDict([("pixel_values", {0: "batch"})])
+        return OrderedDict([("pixel_values", {0: "batch", 1: "num_channels"})])
 
     @property
     def outputs(self) -> Mapping[str, Mapping[int, str]]:

src/transformers/models/resnet/configuration_resnet.py

@@ -105,7 +105,7 @@ class ResNetOnnxConfig(OnnxConfig):
     def inputs(self) -> Mapping[str, Mapping[int, str]]:
         return OrderedDict(
             [
-                ("pixel_values", {0: "batch", 1: "sequence"}),
+                ("pixel_values", {0: "batch", 1: "num_channels"}),
             ]
         )
 

src/transformers/models/vit/configuration_vit.py

@@ -135,7 +135,7 @@ class ViTOnnxConfig(OnnxConfig):
     def inputs(self) -> Mapping[str, Mapping[int, str]]:
         return OrderedDict(
             [
-                ("pixel_values", {0: "batch", 1: "sequence"}),
+                ("pixel_values", {0: "batch", 1: "num_channels"}),
             ]
         )
 

tests/onnx/test_onnx_v2.py

@@ -199,6 +199,7 @@ PYTORCH_EXPORT_MODELS = {
     ("roformer", "junnyu/roformer_chinese_base"),
     ("squeezebert", "squeezebert/squeezebert-uncased"),
     ("mobilebert", "google/mobilebert-uncased"),
+    ("mobilevit", "apple/mobilevit-small"),
     ("xlm", "xlm-clm-ende-1024"),
     ("xlm-roberta", "xlm-roberta-base"),
     ("layoutlm", "microsoft/layoutlm-base-uncased"),