Arm backend: update the pretrained flag (#15739)

cc @freddan80 @per @zingo @oscarandersson8218 @digantdesai

Signed-off-by: Tirui Wu <[email protected]>

Author: tirwu01

examples/models/deit_tiny/model.py

@@ -6,7 +6,6 @@
 import logging
 
 import torch
-from torchvision import transforms
 
 try:
     import timm  # type: ignore
@@ -15,8 +14,6 @@
         "timm package is required for builtin 'deit_tiny'. Install timm."
     ) from e
 
-from timm.data import IMAGENET_INCEPTION_MEAN, IMAGENET_INCEPTION_STD
-
 from ..model_base import EagerModelBase
 
 
@@ -27,16 +24,13 @@ def __init__(self):  # type: ignore[override]
 
     def get_eager_model(self) -> torch.nn.Module:  # type: ignore[override]
         logging.info("Loading timm deit_tiny_patch16_224 model")
-        model = timm.models.deit.deit_tiny_patch16_224(pretrained=False)
-        model.eval()
+        model = timm.models.deit.deit_tiny_patch16_224(pretrained=True)
         logging.info("Loaded timm deit_tiny_patch16_224 model")
         return model
 
     def get_example_inputs(self):  # type: ignore[override]
-        normalize = transforms.Normalize(
-            mean=IMAGENET_INCEPTION_MEAN, std=IMAGENET_INCEPTION_STD
-        )
-        return (normalize(torch.rand((1, 3, 224, 224))),)
+        input_shape = (1, 3, 224, 224)
+        return (torch.randn(input_shape),)
 
 
 __all__ = ["DeiTTinyModel"]