Add stop sequence to text generation pipeline (#18444)

tests/generation/test_generation_utils.py

@@ -37,6 +37,7 @@ if is_torch_available():
         Speech2TextForConditionalGeneration,
         SpeechEncoderDecoderModel,
         VisionEncoderDecoderModel,
+        pipeline,
         top_k_top_p_filtering,
     )
     from transformers.generation_beam_constraints import DisjunctiveConstraint, PhrasalConstraint
@@ -1979,6 +1980,25 @@ class GenerationIntegrationTests(unittest.TestCase):
             [1, 18],
         )
 
+    def test_stop_sequence_stopping_criteria(self):
+
+        prompt = """Hello I believe in"""
+        generator = pipeline("text-generation", model="hf-internal-testing/tiny-random-bart")
+        output = generator(prompt)
+        self.assertEqual(
+            output,
+            [
+                {
+                    "generated_text": (
+                        "Hello I believe in in in number number number number number number number number number"
+                    )
+                }
+            ],
+        )
+
+        output = generator(prompt, stop_sequence=" number")
+        self.assertEqual(output, [{"generated_text": "Hello I believe in in in number"}])
+
     def test_custom_logits_processor(self):
         bart_tokenizer = BartTokenizer.from_pretrained("sshleifer/bart-tiny-random")
         article = """Justin Timberlake and Jessica Biel, welcome to parenthood."""

tests/pipelines/test_pipelines_text_generation.py

@@ -147,6 +147,18 @@ class TextGenerationPipelineTests(unittest.TestCase, metaclass=PipelineTestCaseM
         text_generator = TextGenerationPipeline(model=model, tokenizer=tokenizer)
         return text_generator, ["This is a test", "Another test"]
 
+    def test_stop_sequence_stopping_criteria(self):
+        prompt = """Hello I believe in"""
+        text_generator = pipeline("text-generation", model="hf-internal-testing/tiny-random-gpt2")
+        output = text_generator(prompt)
+        self.assertEqual(
+            output,
+            [{"generated_text": "Hello I believe in fe fe fe fe fe fe fe fe fe fe fe fe"}],
+        )
+
+        output = text_generator(prompt, stop_sequence=" fe")
+        self.assertEqual(output, [{"generated_text": "Hello I believe in fe"}])
+
     def run_pipeline_test(self, text_generator, _):
         model = text_generator.model
         tokenizer = text_generator.tokenizer