Fix length related warnings in speculative decoding (#29585)

* avoid generation length warning

* add tests

* Update src/transformers/generation/candidate_generator.py
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* add tests and minor fixes

* refine `min_new_tokens`

* Update src/transformers/generation/candidate_generator.py
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* add method to prepare length arguments

* add test for min length

* Update src/transformers/generation/candidate_generator.py
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* fix variable naming

* empty commit for tests

* trigger tests (empty)

---------
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

src/transformers/generation/candidate_generator.py

@@ -148,6 +148,11 @@ class AssistedCandidateGenerator(CandidateGenerator):
         self.generation_config.return_dict_in_generate = True
         self.generation_config.output_scores = True
 
+        # avoid unnecessary warnings that min_length is larger than max_new_tokens
+        self.main_model_min_length = self.generation_config.min_length
+        self.generation_config.min_length = 0
+        self.generation_config.min_new_tokens = None
+
     def get_candidates(self, input_ids: torch.LongTensor) -> Tuple[torch.LongTensor, Optional[torch.FloatTensor]]:
         """
         Fetches the candidates to be tried for the current input.
@@ -166,6 +171,7 @@ class AssistedCandidateGenerator(CandidateGenerator):
         # Don't generate more than `max_length - 1` candidates since the target model generates one extra token.
         new_cur_len = input_ids.shape[-1]
         max_new_tokens = min(int(self.num_assistant_tokens), self.generation_config.max_length - new_cur_len - 1)
+        min_new_tokens = max(min(max_new_tokens, self.main_model_min_length - new_cur_len), 0)
         if max_new_tokens == 0:
             return input_ids, None
 
@@ -186,6 +192,7 @@ class AssistedCandidateGenerator(CandidateGenerator):
         # 2. Forecast next N tokens using the assistant model.
         assistant_generation_kwargs = {
             self.input_ids_key: input_ids,
+            "min_new_tokens": min_new_tokens,
             "max_new_tokens": max_new_tokens,
             "generation_config": self.generation_config,
             "logits_processor": self.logits_processor,

src/transformers/generation/utils.py

@@ -1173,6 +1173,56 @@ class GenerationMixin:
                     UserWarning,
                 )
 
+    def _prepare_generated_length(
+        self,
+        generation_config,
+        has_default_max_length,
+        has_default_min_length,
+        model_input_name,
+        input_ids_length,
+        inputs_tensor,
+    ):
+        """Prepared max and min length in generaion configs to avoid clashes between similar attributes"""
+
+        if generation_config.max_new_tokens is not None:
+            if not has_default_max_length and generation_config.max_length is not None:
+                logger.warning(
+                    f"Both `max_new_tokens` (={generation_config.max_new_tokens}) and `max_length`(="
+                    f"{generation_config.max_length}) seem to have been set. `max_new_tokens` will take precedence. "
+                    "Please refer to the documentation for more information. "
+                    "(https://huggingface.co/docs/transformers/main/en/main_classes/text_generation)"
+                )
+            generation_config.max_length = generation_config.max_new_tokens + input_ids_length
+
+        # if both `inputs_embeds` and `input_ids` are passed, we do not correct the length
+        # otherwise we need total length [inputs-embeds-len + new-tokens-len] to not go beyond indicated `max_length``
+        elif (
+            model_input_name == "inputs_embeds"
+            and input_ids_length != inputs_tensor.shape[1]
+            and not self.config.is_encoder_decoder
+        ):
+            generation_config.max_length -= inputs_tensor.shape[1]
+
+        # same for min length
+        if generation_config.min_new_tokens is not None:
+            if not has_default_min_length:
+                logger.warning(
+                    f"Both `min_new_tokens` (={generation_config.min_new_tokens}) and `min_length`(="
+                    f"{generation_config.min_length}) seem to have been set. `min_new_tokens` will take precedence. "
+                    "Please refer to the documentation for more information. "
+                    "(https://huggingface.co/docs/transformers/main/en/main_classes/text_generation)"
+                )
+            generation_config.min_length = generation_config.min_new_tokens + input_ids_length
+
+        elif (
+            model_input_name == "inputs_embeds"
+            and input_ids_length != inputs_tensor.shape[1]
+            and not self.config.is_encoder_decoder
+        ):
+            generation_config.min_length = max(generation_config.min_length - inputs_tensor.shape[1], 0)
+
+        return generation_config
+
     def _prepare_generation_config(
         self, generation_config: GenerationConfig, **kwargs: Dict
     ) -> Tuple[GenerationConfig, Dict]:
@@ -1418,24 +1468,15 @@ class GenerationMixin:
         # 6. Prepare `max_length` depending on other stopping criteria.
         input_ids_length = input_ids.shape[-1]
         has_default_max_length = kwargs.get("max_length") is None and generation_config.max_length is not None
-        if generation_config.max_new_tokens is not None:
-            if not has_default_max_length and generation_config.max_length is not None:
-                logger.warning(
-                    f"Both `max_new_tokens` (={generation_config.max_new_tokens}) and `max_length`(="
-                    f"{generation_config.max_length}) seem to have been set. `max_new_tokens` will take precedence. "
-                    "Please refer to the documentation for more information. "
-                    "(https://huggingface.co/docs/transformers/main/en/main_classes/text_generation)"
-                )
-            generation_config.max_length = generation_config.max_new_tokens + input_ids_length
-
-        # otherwise the total length [inputs-embeds-len + new-tokens-len] will go beyond indicated `max_length``
-        elif (
-            model_input_name == "inputs_embeds"
-            and inputs_tensor.shape[:-1] != input_ids.shape
-            and not self.config.is_encoder_decoder
-        ):
-            generation_config.max_length -= inputs_tensor.shape[1]
-            generation_config.min_length = max(generation_config.min_length - inputs_tensor.shape[1], 0)
+        has_default_min_length = kwargs.get("min_length") is None and generation_config.min_length is not None
+        generation_config = self._prepare_generated_length(
+            generation_config=generation_config,
+            has_default_max_length=has_default_max_length,
+            has_default_min_length=has_default_min_length,
+            model_input_name=model_input_name,
+            inputs_tensor=inputs_tensor,
+            input_ids_length=input_ids_length,
+        )
 
         if generation_config.cache_implementation in NEED_SETUP_CACHE_CLASSES_MAPPING:
             if generation_config.cache_implementation == "static":
@@ -1511,7 +1552,7 @@ class GenerationMixin:
             )
 
             # 12. run assisted generate
-            result = self.assisted_decoding(
+            result = self._assisted_decoding(
                 input_ids,
                 candidate_generator=candidate_generator,
                 do_sample=generation_config.do_sample,

tests/generation/test_utils.py

@@ -1977,6 +1977,20 @@ class GenerationIntegrationTests(unittest.TestCase, GenerationIntegrationTestsMi
         out_gen_embeds = model.generate(inputs_embeds=inputs_embeds, max_length=max_length)
         self.assertEqual(out_gen.shape[-1], input_len + out_gen_embeds.shape[-1])
 
+    def test_min_length_if_input_embeds(self):
+        # PT-only test: TF doesn't have StoppingCriteria
+        article = "Today a dragon flew over Paris."
+        model = AutoModelForCausalLM.from_pretrained("hf-internal-testing/tiny-random-gpt2").to(torch_device)
+        tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-gpt2")
+        input_ids = tokenizer(article, return_tensors="pt").input_ids.to(torch_device)
+        inputs_embeds = model.get_input_embeddings()(input_ids)
+
+        min_length = 10
+        input_len = input_ids.shape[-1]
+        out_gen = model.generate(input_ids=input_ids, min_length=min_length)
+        out_gen_embeds = model.generate(inputs_embeds=inputs_embeds, min_length=min_length)
+        self.assertEqual(out_gen.shape[-1], input_len + out_gen_embeds.shape[-1])
+
     def test_custom_stopping_criteria_overload_error(self):
         # PT-only test: TF doesn't have StoppingCriteria
         article = """Justin Timberlake and Jessica Biel, welcome to parenthood."""
@@ -2539,6 +2553,56 @@ class GenerationIntegrationTests(unittest.TestCase, GenerationIntegrationTestsMi
             model.generate(input_ids)
             self.assertEqual(len(warning_list), 0)
 
+    def test_length_warning_assisted_generation(self):
+        # PT-only test: TF doesn't support assisted decoding yet.
+        model = AutoModelForCausalLM.from_pretrained("hf-internal-testing/tiny-random-gpt2").to(torch_device)
+        assistant = AutoModelForCausalLM.from_pretrained("hf-internal-testing/tiny-random-gpt2").to(torch_device)
+        tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-gpt2")
+        model.config.pad_token_id = tokenizer.eos_token_id
+        assistant.config.pad_token_id = tokenizer.eos_token_id
+
+        text = "Hello world"
+        tokenized_inputs = tokenizer([text], return_tensors="pt")
+        input_ids = tokenized_inputs.input_ids.to(torch_device)
+
+        # This should not raise any warning that min length is not feasible in candidate generation
+        with warnings.catch_warnings(record=True) as warning_list:
+            model.generate(
+                input_ids,
+                assistant_model=assistant,
+                min_new_tokens=10,
+                max_length=20,
+            )
+            self.assertEqual(len(warning_list), 0)
+
+    def test_generated_length_assisted_generation(self):
+        # PT-only test: TF doesn't support assisted decoding yet.
+        model = AutoModelForCausalLM.from_pretrained("hf-internal-testing/tiny-random-gpt2").to(torch_device)
+        assistant = AutoModelForCausalLM.from_pretrained("hf-internal-testing/tiny-random-gpt2").to(torch_device)
+        tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-gpt2")
+        model.config.pad_token_id = tokenizer.eos_token_id
+        assistant.config.pad_token_id = tokenizer.eos_token_id
+
+        text = "Hello world"
+        tokenized_inputs = tokenizer([text], return_tensors="pt")
+        input_ids = tokenized_inputs.input_ids.to(torch_device)
+        input_length = input_ids.shape[-1]
+
+        out = model.generate(
+            input_ids,
+            assistant_model=assistant,
+            min_new_tokens=10,
+            max_new_tokens=20,
+        )
+        self.assertTrue((10 + input_length) <= out.shape[-1] <= (20 + input_length))
+
+        out = model.generate(
+            input_ids,
+            assistant_model=assistant,
+            min_new_tokens=10,
+        )
+        self.assertTrue((input_length + 10) <= out.shape[-1] <= 20)
+
     def test_model_kwarg_assisted_decoding_decoder_only(self):
         # PT-only test: TF doesn't support assisted decoding yet.
         model = AutoModelForCausalLM.from_pretrained("hf-internal-testing/tiny-random-gpt2").to(torch_device)