Shorten the conversation tests for speed + fixing position overflows

tests/models/blenderbot/test_modeling_blenderbot.py

@@ -85,7 +85,7 @@ def __init__(
         hidden_act="gelu",
         hidden_dropout_prob=0.1,
         attention_probs_dropout_prob=0.1,
-        max_position_embeddings=20,
+        max_position_embeddings=50,
         eos_token_id=2,
         pad_token_id=1,
         bos_token_id=0,

tests/models/blenderbot/test_modeling_flax_blenderbot.py

@@ -87,7 +87,7 @@ def __init__(
         hidden_act="gelu",
         hidden_dropout_prob=0.1,
         attention_probs_dropout_prob=0.1,
-        max_position_embeddings=32,
+        max_position_embeddings=50,
         eos_token_id=2,
         pad_token_id=1,
         bos_token_id=0,

tests/models/blenderbot/test_modeling_tf_blenderbot.py

@@ -53,7 +53,7 @@ def __init__(
         intermediate_size=37,
         hidden_dropout_prob=0.1,
         attention_probs_dropout_prob=0.1,
-        max_position_embeddings=20,
+        max_position_embeddings=50,
         eos_token_id=2,
         pad_token_id=1,
         bos_token_id=0,

tests/models/blenderbot_small/test_modeling_blenderbot_small.py

@@ -85,7 +85,7 @@ def __init__(
         hidden_act="gelu",
         hidden_dropout_prob=0.1,
         attention_probs_dropout_prob=0.1,
-        max_position_embeddings=20,
+        max_position_embeddings=50,
         eos_token_id=2,
         pad_token_id=1,
         bos_token_id=0,
@@ -244,9 +244,6 @@ def is_pipeline_test_to_skip(
     ):
         if pipeline_test_casse_name == "TextGenerationPipelineTests":
             return True
-        # TODO @Rocketnight1 to fix
-        if pipeline_test_casse_name == "ConversationalPipelineTests":
-            return True
         return False
 
     def setUp(self):
@@ -284,6 +281,10 @@ def test_generate_fp16(self):
         model.generate(input_ids, attention_mask=attention_mask)
         model.generate(num_beams=4, do_sample=True, early_stopping=False, num_return_sequences=3)
 
+    @unittest.skip("Tiny random model has too few position embeddings for this.")
+    def test_pipeline_conversational(self):
+        pass
+
 
 def assert_tensors_close(a, b, atol=1e-12, prefix=""):
     """If tensors have different shapes, different values or a and b are not both tensors, raise a nice Assertion error."""

tests/models/blenderbot_small/test_modeling_flax_blenderbot_small.py

@@ -86,7 +86,7 @@ def __init__(
         hidden_act="gelu",
         hidden_dropout_prob=0.1,
         attention_probs_dropout_prob=0.1,
-        max_position_embeddings=32,
+        max_position_embeddings=50,
         eos_token_id=2,
         pad_token_id=1,
         bos_token_id=0,
@@ -397,3 +397,7 @@ def test_model_from_pretrained(self):
             input_ids = np.ones((1, 1)) * model.config.eos_token_id
             outputs = model(input_ids)
             self.assertIsNotNone(outputs)
+
+    @unittest.skip("Tiny random model has too few position embeddings for this.")
+    def test_pipeline_conversational(self):
+        pass

tests/models/blenderbot_small/test_modeling_tf_blenderbot_small.py

@@ -53,7 +53,7 @@ def __init__(
         intermediate_size=37,
         hidden_dropout_prob=0.1,
         attention_probs_dropout_prob=0.1,
-        max_position_embeddings=20,
+        max_position_embeddings=50,
         eos_token_id=2,
         pad_token_id=1,
         bos_token_id=0,
@@ -209,14 +209,9 @@ def test_decoder_model_past_large_inputs(self):
         config_and_inputs = self.model_tester.prepare_config_and_inputs_for_common()
         self.model_tester.check_decoder_model_past_large_inputs(*config_and_inputs)
 
-    # TODO: Fix the failed tests when this model gets more usage
-    def is_pipeline_test_to_skip(
-        self, pipeline_test_casse_name, config_class, model_architecture, tokenizer_name, processor_name
-    ):
-        # TODO @Rocketnight1 to fix
-        if pipeline_test_casse_name == "ConversationalPipelineTests":
-            return True
-        return False
+    @unittest.skip("Tiny random model has too few position embeddings for this.")
+    def test_pipeline_conversational(self):
+        pass
 
 
 @require_tokenizers

tests/pipelines/test_pipelines_conversational.py

@@ -77,14 +77,14 @@ def get_test_pipeline(self, model, tokenizer, processor):
 
     def run_pipeline_test(self, conversation_agent, _):
         # Simple
-        outputs = conversation_agent(Conversation("Hi there!"), max_new_tokens=20)
+        outputs = conversation_agent(Conversation("Hi there!"), max_new_tokens=5)
         self.assertEqual(
             outputs,
             Conversation([{"role": "user", "content": "Hi there!"}, {"role": "assistant", "content": ANY(str)}]),
         )
 
         # Single list
-        outputs = conversation_agent([Conversation("Hi there!")], max_new_tokens=20)
+        outputs = conversation_agent([Conversation("Hi there!")], max_new_tokens=5)
         self.assertEqual(
             outputs,
             Conversation([{"role": "user", "content": "Hi there!"}, {"role": "assistant", "content": ANY(str)}]),
@@ -96,7 +96,7 @@ def run_pipeline_test(self, conversation_agent, _):
         self.assertEqual(len(conversation_1), 1)
         self.assertEqual(len(conversation_2), 1)
 
-        outputs = conversation_agent([conversation_1, conversation_2], max_new_tokens=20)
+        outputs = conversation_agent([conversation_1, conversation_2], max_new_tokens=5)
         self.assertEqual(outputs, [conversation_1, conversation_2])
         self.assertEqual(
             outputs,
@@ -118,7 +118,7 @@ def run_pipeline_test(self, conversation_agent, _):
 
         # One conversation with history
         conversation_2.add_message({"role": "user", "content": "Why do you recommend it?"})
-        outputs = conversation_agent(conversation_2, max_new_tokens=20)
+        outputs = conversation_agent(conversation_2, max_new_tokens=5)
         self.assertEqual(outputs, conversation_2)
         self.assertEqual(
             outputs,

tests/test_pipeline_mixin.py

@@ -313,7 +313,7 @@ def data(n):
 
             out = []
             if task == "conversational":
-                for item in pipeline(data(10), batch_size=4, max_new_tokens=20):
+                for item in pipeline(data(10), batch_size=4, max_new_tokens=5):
                     out.append(item)
             else:
                 for item in pipeline(data(10), batch_size=4):