Fix tests for vision models (#35654)

* Trigger tests

* [run-slow] beit, detr, dinov2, vit, textnet

* Fix BEiT interpolate_pos_encoding

* Fix DETR test

* Update DINOv2 test

* Fix textnet

* Fix vit

* Fix DPT

* fix data2vec test

* Fix textnet test

* Update interpolation check

* Fix ZoeDepth tests

* Update interpolate embeddings for BEiT

* Apply suggestions from code review

tests/models/detr/test_modeling_detr.py

@@ -684,7 +684,12 @@ class DetrModelIntegrationTestsTimmBackbone(unittest.TestCase):
         self.assertTrue(results["segmentation"].shape, expected_shape)
         torch.testing.assert_close(results["segmentation"][:3, :3], expected_slice_segmentation, rtol=1e-4, atol=1e-4)
         self.assertTrue(len(results["segments_info"]), expected_number_of_segments)
-        self.assertDictEqual(results["segments_info"][0], expected_first_segment)
+
+        predicted_first_segment = results["segments_info"][0]
+        self.assertEqual(predicted_first_segment["id"], expected_first_segment["id"])
+        self.assertEqual(predicted_first_segment["label_id"], expected_first_segment["label_id"])
+        self.assertEqual(predicted_first_segment["was_fused"], expected_first_segment["was_fused"])
+        self.assertAlmostEqual(predicted_first_segment["score"], expected_first_segment["score"], places=3)
 
 
 @require_vision