[Bugfix] Whisper fix number of allocated CrossAttn blocks per-request (#30772)

Signed-off-by: NickLucche <nlucches@redhat.com>

vllm/v1/core/sched/scheduler.py

@@ -187,6 +187,12 @@ class Scheduler(SchedulerInterface):
             if self.is_encoder_decoder
             else EncoderCacheManager(cache_size=encoder_cache_size)
         )
+        # For encoder-decoder models, allocate the maximum number of tokens for Cross
+        # Attn blocks, as for Whisper its input is always padded to the maximum length.
+        # TODO (NickLucche): Generalize to models with variable-length encoder inputs.
+        self._num_encoder_max_input_tokens = (
+            MULTIMODAL_REGISTRY.get_encdec_max_encoder_len(vllm_config.model_config)
+        )
 
         speculative_config = vllm_config.speculative_config
         self.use_eagle = False
@@ -568,17 +574,11 @@ class Scheduler(SchedulerInterface):
                     0 if request.num_computed_tokens == 0 else self.num_lookahead_tokens
                 )
 
-                # Determine if we need to allocate cross-attention blocks.
-                if self.is_encoder_decoder and request.has_encoder_inputs:
-                    # TODO(russellb): For Whisper, we know that the input is
-                    # always padded to the maximum length. If we support other
-                    # encoder-decoder models, this will need to be updated if we
-                    # want to only allocate what is needed.
                 num_encoder_tokens = (
-                        self.scheduler_config.max_num_encoder_input_tokens
+                    self._num_encoder_max_input_tokens
+                    if self.is_encoder_decoder and request.has_encoder_inputs
+                    else 0
                 )
-                else:
-                    num_encoder_tokens = 0
 
                 new_blocks = self.kv_cache_manager.allocate_slots(
                     request,