Allow FP16 or other precision inference for Pipelines (#31342)

* cast image features to model.dtype where needed to support FP16 or other precision in pipelines

* Update src/transformers/pipelines/image_feature_extraction.py

Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* Use .to instead

* Add FP16 pipeline support for zeroshot audio classification

* Remove unused torch imports

* Add docs on FP16 pipeline

* Remove unused import

* Add FP16 tests to pipeline mixin

* Add fp16 placeholder for mask_generation pipeline test

* Add FP16 tests for all pipelines

* Fix formatting

* Remove torch_dtype arg from is_pipeline_test_to_skip*

* Fix format

* trigger ci

---------

Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

tests/pipelines/test_pipelines_image_feature_extraction.py

@@ -157,7 +157,7 @@ class ImageFeatureExtractionPipelineTests(unittest.TestCase):
         outputs = feature_extractor(img, return_tensors=True)
         self.assertTrue(tf.is_tensor(outputs))
 
-    def get_test_pipeline(self, model, tokenizer, processor):
+    def get_test_pipeline(self, model, tokenizer, processor, torch_dtype="float32"):
         if processor is None:
             self.skipTest(reason="No image processor")
 
@@ -175,7 +175,9 @@ class ImageFeatureExtractionPipelineTests(unittest.TestCase):
                 """
             )
 
-        feature_extractor = ImageFeatureExtractionPipeline(model=model, image_processor=processor)
+        feature_extractor = ImageFeatureExtractionPipeline(
+            model=model, image_processor=processor, torch_dtype=torch_dtype
+        )
         img = prepare_img()
         return feature_extractor, [img, img]