[Misc] Add quantization config support for speculative model. (vllm-p…

…roject#7343)

Signed-off-by: Alvant <alvasian@yandex.ru>

tests/spec_decode/e2e/test_integration.py

@@ -42,3 +42,51 @@ def test_spec_decode_cuda_graph(baseline_llm_generator, test_llm_generator,
         max_output_len=output_len,
         force_output_len=True,
     )
+
+
+@pytest.mark.parametrize(
+    "common_llm_kwargs",
+    [{
+        "model": "JackFram/llama-160m",
+
+        # Skip cuda graph recording for fast test.
+        "enforce_eager": True,
+
+        # Required for spec decode.
+        "use_v2_block_manager": True,
+    }])
+@pytest.mark.parametrize("per_test_common_llm_kwargs", [
+    {
+        "speculative_model": "LnL-AI/TinyLlama-1.1B-Chat-v1.0-GPTQ-4bit",
+        "num_speculative_tokens": 5,
+    },
+])
+@pytest.mark.parametrize(
+    "test_llm_kwargs",
+    [
+        # Explicitly specify draft model quantization
+        {
+            "speculative_model_quantization": "gptq",
+        },
+        # Explicitly specify GPTQ-based draft model to use marlin quantization
+        {
+            "speculative_model_quantization": "marlin",
+        },
+        # Not explicitly specify draft model quantization
+        {
+            "speculative_model_quantization": None,
+        },
+    ])
+@pytest.mark.parametrize("baseline_llm_kwargs", [{}])
+@pytest.mark.parametrize("batch_size", [2])
+@pytest.mark.parametrize("seed", [1])
+def test_speculative_model_quantization_config(baseline_llm_generator,
+                                               test_llm_generator,
+                                               batch_size: int):
+    """Verify spec decode works well with draft model quantization configs.
+    """
+    run_greedy_equality_correctness_test(baseline_llm_generator,
+                                         test_llm_generator,
+                                         batch_size,
+                                         max_output_len=32,
+                                         force_output_len=True)

vllm/config.py

@@ -961,6 +961,7 @@ def maybe_create_spec_config(
         target_parallel_config: ParallelConfig,
         target_dtype: str,
         speculative_model: Optional[str],
+        speculative_model_quantization: Optional[str],
         speculative_draft_tensor_parallel_size: Optional[int],
         num_speculative_tokens: Optional[int],
         speculative_max_model_len: Optional[int],
@@ -989,6 +990,9 @@ def maybe_create_spec_config(
             target_dtype (str): The data type used for the target model.
             speculative_model (Optional[str]): The name of the speculative
                 model, if provided.
+            speculative_model_quantization (Optional[str]): Quantization method
+                that was used to quantize the speculative model weights. If
+                None, we assume the model weights are not quantized.
             speculative_draft_tensor_parallel_size (Optional[int]): The degree
                 of the tensor parallelism for the draft model.
             num_speculative_tokens (Optional[int]): The number of speculative
@@ -1056,11 +1060,11 @@ def maybe_create_spec_config(
                 "Speculative decoding requires usage of the V2 "
                 "block manager. Enable it with --use-v2-block-manager.")
 
-        # TODO: The user should be able to specify revision/quantization/max
-        # model len for the draft model. It is not currently supported.
+        # TODO: The user should be able to specify revision/max model len
+        # for the draft model. It is not currently supported.
         draft_revision = None
         draft_code_revision = None
-        draft_quantization = None
+        draft_quantization = speculative_model_quantization
 
         if speculative_model == "[ngram]":
             if ngram_prompt_lookup_min is None:
@@ -1217,7 +1221,7 @@ def create_draft_parallel_config(
         elif speculative_draft_tensor_parallel_size != 1:
             # TODO(wooyeon): allow tp values larger than 1
             raise ValueError(
-                f"{speculative_draft_tensor_parallel_size=} cannot be"
+                f"{speculative_draft_tensor_parallel_size=} cannot be "
                 f"other value than 1")
 
         draft_parallel_config = ParallelConfig(

vllm/engine/arg_utils.py

@@ -129,6 +129,7 @@ class EngineArgs:
     guided_decoding_backend: str = 'outlines'
     # Speculative decoding configuration.
     speculative_model: Optional[str] = None
+    speculative_model_quantization: Optional[str] = None
     speculative_draft_tensor_parallel_size: Optional[int] = None
     num_speculative_tokens: Optional[int] = None
     speculative_max_model_len: Optional[int] = None
@@ -571,6 +572,18 @@ def add_cli_args(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
             default=EngineArgs.speculative_model,
             help=
             'The name of the draft model to be used in speculative decoding.')
+        # Quantization settings for speculative model.
+        parser.add_argument(
+            '--speculative-model-quantization',
+            type=nullable_str,
+            choices=[*QUANTIZATION_METHODS, None],
+            default=EngineArgs.speculative_model_quantization,
+            help='Method used to quantize the weights of speculative model.'
+            'If None, we first check the `quantization_config` '
+            'attribute in the model config file. If that is '
+            'None, we assume the model weights are not '
+            'quantized and use `dtype` to determine the data '
+            'type of the weights.')
         parser.add_argument(
             '--num-speculative-tokens',
             type=int,
@@ -844,6 +857,8 @@ def create_engine_config(self, ) -> EngineConfig:
             target_parallel_config=parallel_config,
             target_dtype=self.dtype,
             speculative_model=self.speculative_model,
+            speculative_model_quantization = \
+                self.speculative_model_quantization,
             speculative_draft_tensor_parallel_size = \
                 self.speculative_draft_tensor_parallel_size,
             num_speculative_tokens=self.num_speculative_tokens,