Replace as_target context managers by direct calls (#18325)

* Preliminary work on tokenizers

* Quality + fix tests

* Treat processors

* Fix pad

* Remove all uses of  in tests, docs and examples

* Replace all as_target_tokenizer

* Fix tests

* Fix quality

* Update examples/flax/image-captioning/run_image_captioning_flax.py

Co-authored-by: amyeroberts <amy@huggingface.co>

* Style

Co-authored-by: amyeroberts <amy@huggingface.co>

examples/research_projects/xtreme-s/run_xtreme_s.py

@@ -349,13 +349,12 @@ class SpeechDataCollatorWithPadding:
 
         if self.pad_labels:
             label_features = [{"input_ids": feature["labels"]} for feature in features]
-            with self.processor.as_target_processor():
-                labels_batch = self.processor.pad(
-                    label_features,
-                    padding=self.padding,
-                    pad_to_multiple_of=self.pad_to_multiple_of_labels,
-                    return_tensors="pt",
-                )
+            labels_batch = self.processor.pad(
+                labels=label_features,
+                padding=self.padding,
+                pad_to_multiple_of=self.pad_to_multiple_of_labels,
+                return_tensors="pt",
+            )
 
             # replace padding with -100 to ignore loss correctly
             labels = labels_batch["input_ids"].masked_fill(labels_batch.attention_mask.ne(1), -100)