Automatically configure max_num_batched_tokens (#1198)

Author: WoosukKwon

vllm/config.py

@@ -266,11 +266,36 @@ class SchedulerConfig:
             and generated text).
     """
 
-    def __init__(self, max_num_batched_tokens: int, max_num_seqs: int,
-                 max_model_len: int) -> None:
-        self.max_num_batched_tokens = max_num_batched_tokens
+    def __init__(
+        self,
+        max_num_batched_tokens: Optional[int],
+        max_num_seqs: int,
+        max_model_len: int,
+    ) -> None:
+        if max_num_batched_tokens is not None:
+            self.max_num_batched_tokens = max_num_batched_tokens
+        else:
+            # If max_model_len is too short, use 2048 as the default value for
+            # higher throughput.
+            self.max_num_batched_tokens = max(max_model_len, 2048)
         self.max_num_seqs = max_num_seqs
         self.max_model_len = max_model_len
+        self._verify_args()
+
+    def _verify_args(self) -> None:
+        if self.max_num_batched_tokens < self.max_model_len:
+            raise ValueError(
+                f"max_num_batched_tokens ({self.max_num_batched_tokens}) is "
+                f"smaller than max_model_len ({self.max_model_len}). "
+                "This effectively limits the maximum sequence length to "
+                "max_num_batched_tokens and makes vLLM reject longer "
+                "sequences. Please increase max_num_batched_tokens or "
+                "decrease max_model_len.")
+        if self.max_num_batched_tokens < self.max_num_seqs:
+            raise ValueError(
+                f"max_num_batched_tokens ({self.max_num_batched_tokens}) must "
+                "be greater than or equal to max_num_seqs "
+                f"({self.max_num_seqs}).")
 
 
 _STR_DTYPE_TO_TORCH_DTYPE = {
@@ -350,14 +375,14 @@ def _get_and_verify_max_len(
         max_len_key = getattr(hf_config, key, None)
         if max_len_key is not None:
             derived_max_model_len = min(derived_max_model_len, max_len_key)
+    if derived_max_model_len == float("inf"):
+        raise ValueError(
+            "The model's config.json must contain one of the following keys "
+            "to determine the original maximum length of the model: "
+            f"{possible_keys}")
 
     rope_scaling = getattr(hf_config, "rope_scaling", None)
     if rope_scaling is not None:
-        if derived_max_model_len == float("inf"):
-            raise ValueError(
-                "When using rope_scaling, the model's config.json must "
-                "contain one of the following keys to determine the original "
-                f"maximum length of the model: {possible_keys}")
         assert "factor" in rope_scaling
         scaling_factor = rope_scaling["factor"]
         derived_max_model_len *= scaling_factor
@@ -371,4 +396,4 @@ def _get_and_verify_max_len(
             " in model's config.json). This may lead to incorrect model "
             "outputs or CUDA errors. Make sure the value is correct and "
             "within the model context size.")
-    return max_model_len
+    return int(max_model_len)

vllm/engine/arg_utils.py

@@ -25,7 +25,7 @@ class EngineArgs:
     block_size: int = 16
     swap_space: int = 4  # GiB
     gpu_memory_utilization: float = 0.90
-    max_num_batched_tokens: int = 2560
+    max_num_batched_tokens: Optional[int] = None
     max_num_seqs: int = 256
     disable_log_stats: bool = False
     revision: Optional[str] = None
@@ -34,7 +34,6 @@ class EngineArgs:
     def __post_init__(self):
         if self.tokenizer is None:
             self.tokenizer = self.model
-        self.max_num_seqs = min(self.max_num_seqs, self.max_num_batched_tokens)
 
     @staticmethod
     def add_cli_args(