diff --git a/tests/models/mistral/test_modeling_mistral.py b/tests/models/mistral/test_modeling_mistral.py index cd6459c170dc..1723d55afc8a 100644 --- a/tests/models/mistral/test_modeling_mistral.py +++ b/tests/models/mistral/test_modeling_mistral.py @@ -254,7 +254,7 @@ def test_model_7b_long_prompt_sdpa(self): @slow def test_speculative_generation(self): - EXPECTED_TEXT_COMPLETION = "My favourite condiment is 100% Sriracha. I love it on everything. I have it on my" + EXPECTED_TEXT_COMPLETION = "My favourite condiment is 100% ketchup. I’m not a fan of mustard, relish" prompt = "My favourite condiment is " tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-v0.1", use_fast=False) model = MistralForCausalLM.from_pretrained("mistralai/Mistral-7B-v0.1", device_map="auto", dtype=torch.float16) diff --git a/tests/models/qwen2/test_modeling_qwen2.py b/tests/models/qwen2/test_modeling_qwen2.py index d5c207fb9b8c..d4cf34fbbca2 100644 --- a/tests/models/qwen2/test_modeling_qwen2.py +++ b/tests/models/qwen2/test_modeling_qwen2.py @@ -207,7 +207,7 @@ def test_model_450m_long_prompt_sdpa(self): @slow def test_speculative_generation(self): EXPECTED_TEXT_COMPLETION = ( - "My favourite condiment is 100% natural honey, and I always like to use it in my recipes. I love" + "My favourite condiment is 100% natural and organic, and I love to use it to make my own sauces." ) prompt = "My favourite condiment is " tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2-7B", use_fast=False) diff --git a/tests/models/qwen2_moe/test_modeling_qwen2_moe.py b/tests/models/qwen2_moe/test_modeling_qwen2_moe.py index 02eb4dfb4643..db3be5ac7e20 100644 --- a/tests/models/qwen2_moe/test_modeling_qwen2_moe.py +++ b/tests/models/qwen2_moe/test_modeling_qwen2_moe.py @@ -252,7 +252,7 @@ def test_model_a2_7b_long_prompt_sdpa(self): @slow def test_speculative_generation(self): EXPECTED_TEXT_COMPLETION = ( - "To be or not to be, that is the question. Whether 'tis nobler in the mind to suffer the sl" + "To be or not to be, that is the question: Whether 'tis nobler in the mind to suffer The sl" ) prompt = "To be or not to" tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen1.5-MoE-A2.7B", use_fast=False) diff --git a/tests/models/qwen3/test_modeling_qwen3.py b/tests/models/qwen3/test_modeling_qwen3.py index 169d7ee784a0..ba937656d3a6 100644 --- a/tests/models/qwen3/test_modeling_qwen3.py +++ b/tests/models/qwen3/test_modeling_qwen3.py @@ -199,7 +199,7 @@ def test_speculative_generation(self): { ("xpu", 3): "My favourite condiment is 100% peanut butter. I love it so much that I can't help but use it", ("cuda", 7): "My favourite condiment is 100% natural. It's a little spicy and a little sweet, but it's the", - ("cuda", 8): "My favourite condiment is 100% peanut butter. I love it so much that I can't help but use it", + ("cuda", 8): "My favourite condiment is 100% beef, 100% beef, 100% beef.", } ) # fmt: skip EXPECTED_TEXT_COMPLETION = EXPECTED_TEXT_COMPLETIONS.get_expectation()