[testing] slow tests should be marked as slow (#7895)

* slow tests should be slow

* exception note

* style

* integrate LysandreJik's notes with some expansions

* Apply suggestions from code review

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* another slow test

* fix link, and prose

* clarify.

* note from Sam

* typo

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

tests/test_modeling_marian.py

@@ -295,6 +295,7 @@ class TestMarian_en_ROMANCE(MarianIntegrationTest):
         with self.assertRaises(ValueError):
             self.tokenizer.prepare_seq2seq_batch([""])
 
+    @slow
     def test_pipeline(self):
         device = 0 if torch_device == "cuda" else -1
         pipeline = TranslationPipeline(self.model, self.tokenizer, framework="pt", device=device)

tests/test_pipelines.py

@@ -379,6 +379,7 @@ class MonoColumnInputTestCase(unittest.TestCase):
 
     @require_torch
     @require_tokenizers
+    @slow
     def test_torch_translation(self):
         invalid_inputs = [4, "<mask>"]
         mandatory_keys = ["translation_text"]

tests/test_tokenization_auto.py

@@ -34,12 +34,13 @@ from transformers.testing_utils import (
     DUMMY_UNKWOWN_IDENTIFIER,
     SMALL_MODEL_IDENTIFIER,
     require_tokenizers,
+    slow,
 )
 from transformers.tokenization_auto import TOKENIZER_MAPPING
 
 
 class AutoTokenizerTest(unittest.TestCase):
-    # @slow
+    @slow
     def test_tokenizer_from_pretrained(self):
         for model_name in (x for x in BERT_PRETRAINED_CONFIG_ARCHIVE_MAP.keys() if "japanese" not in x):
             tokenizer = AutoTokenizer.from_pretrained(model_name)