[Core] Factor out common logic for MM budget calculation (#22228)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Cyrus Leung
2025-08-04 23:54:55 -07:00
committed by GitHub
parent e79a12fc3a
commit 811ac13d03
3 changed files with 299 additions and 216 deletions
+105 -112
View File
@@ -36,7 +36,8 @@ from vllm.model_executor.models.interfaces import (is_mixture_of_experts,
from vllm.model_executor.models.interfaces_base import (
VllmModelForPooling, is_pooling_model, is_text_generation_model)
from vllm.multimodal import MULTIMODAL_REGISTRY
from vllm.multimodal.inputs import MultiModalKwargs, PlaceholderRange
from vllm.multimodal.inputs import (BatchedTensorInputs, MultiModalKwargs,
PlaceholderRange)
from vllm.multimodal.utils import group_mm_inputs_by_modality
from vllm.pooling_params import PoolingParams
from vllm.sampling_params import SamplingType
@@ -51,7 +52,6 @@ from vllm.v1.attention.backends.utils import (
make_kv_sharing_fast_prefill_attention_metadata,
make_local_attention_virtual_batches,
reorder_batch_to_split_decodes_and_prefills)
from vllm.v1.core.encoder_cache_manager import compute_encoder_budget
from vllm.v1.kv_cache_interface import (AttentionSpec,
ChunkedLocalAttentionSpec,
FullAttentionSpec, KVCacheConfig,
@@ -73,7 +73,7 @@ from vllm.v1.worker.kv_connector_model_runner_mixin import (
from vllm.v1.worker.lora_model_runner_mixin import LoRAModelRunnerMixin
from ..sample.logits_processor import LogitsProcessorManager
from .utils import (bind_kv_cache, gather_mm_placeholders,
from .utils import (MultiModalBudget, bind_kv_cache, gather_mm_placeholders,
initialize_kv_cache_for_kv_sharing,
sanity_check_mm_encoder_outputs, scatter_mm_placeholders)
@@ -148,14 +148,6 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
self.mm_registry = MULTIMODAL_REGISTRY
self.uses_mrope = model_config.uses_mrope
encoder_compute_budget, encoder_cache_size = compute_encoder_budget(
model_config=model_config,
scheduler_config=scheduler_config,
mm_registry=self.mm_registry,
)
self.max_num_encoder_input_tokens = encoder_compute_budget
self.encoder_cache_size = encoder_cache_size
# Sampler
self.sampler = Sampler(logprobs_mode=self.model_config.logprobs_mode)
@@ -330,6 +322,14 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
self.kv_sharing_fast_prefill_logits_indices = torch.zeros(
self.max_num_tokens, dtype=torch.int32, device=self.device)
self.mm_budget = (MultiModalBudget(
self.model_config,
self.scheduler_config,
self.mm_registry,
max_model_len=self.max_model_len,
max_num_reqs=self.max_num_reqs,
) if self.is_multimodal_model else None)
self.reorder_batch_threshold: Optional[int] = None
def _may_reorder_batch(self, scheduler_output: "SchedulerOutput") -> None:
@@ -578,37 +578,33 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
# Refresh batch metadata with any pending updates.
self.input_batch.refresh_metadata()
def _init_model_kwargs_for_multimodal_model(
def _extract_mm_kwargs(
self,
scheduler_output: Optional["SchedulerOutput"] = None,
num_reqs: int = -1,
) -> dict[str, Any]:
model_kwargs: dict[str, Any] = {}
if self.is_multimodal_raw_input_supported:
# This model requires the raw multimodal data in input.
scheduler_output: "SchedulerOutput",
) -> BatchedTensorInputs:
if self.is_multimodal_raw_input_supported: # noqa: SIM102
if scheduler_output:
multi_modal_kwargs_list = []
multi_modal_kwargs_list = list[MultiModalKwargs]()
for req in scheduler_output.scheduled_new_reqs:
req_mm_inputs = req.mm_inputs
if not isinstance(req_mm_inputs, list):
req_mm_inputs = list(req_mm_inputs)
multi_modal_kwargs_list.extend(req_mm_inputs)
multi_modal_kwargs = MultiModalKwargs.batch(
multi_modal_kwargs_list)
else:
# The only case where SchedulerOutput is None is for
# a dummy run let's get some dummy data.
dummy_data = [
self.mm_registry.get_decoder_dummy_data(
model_config=self.model_config,
seq_len=1).multi_modal_data for i in range(num_reqs)
]
multi_modal_kwargs = MultiModalKwargs.batch(dummy_data)
model_kwargs.update(multi_modal_kwargs)
return MultiModalKwargs.batch(multi_modal_kwargs_list)
return model_kwargs
return {}
def _dummy_mm_kwargs(self, num_seqs: int) -> BatchedTensorInputs:
if self.is_multimodal_raw_input_supported:
mm_budget = self.mm_budget
assert mm_budget is not None
dummy_modality, _ = mm_budget.get_modality_with_max_tokens()
return self._get_mm_dummy_batch(dummy_modality, num_seqs)
return {}
def _get_cumsum_and_arange(
self,
@@ -1517,19 +1513,18 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
# NOTE(woosuk): To unify token ids and soft tokens (vision
# embeddings), we always use embeddings (rather than token ids)
# as input to the multimodal model, even when the input is text.
input_ids = self.input_ids[:num_scheduled_tokens]
model_kwargs = self._init_model_kwargs_for_multimodal_model(
scheduler_output=scheduler_output)
inputs_embeds = self.model.get_input_embeddings(
input_ids=input_ids,
inputs_embeds_scheduled = self.model.get_input_embeddings(
input_ids=self.input_ids[:num_scheduled_tokens],
multimodal_embeddings=mm_embeds or None,
)
# TODO(woosuk): Avoid the copy. Optimize.
self.inputs_embeds[:num_scheduled_tokens].copy_(inputs_embeds)
inputs_embeds = self.inputs_embeds[:num_input_tokens]
self.inputs_embeds[:num_scheduled_tokens].copy_(
inputs_embeds_scheduled)
input_ids = None
inputs_embeds = self.inputs_embeds[:num_input_tokens]
model_mm_kwargs = self._extract_mm_kwargs(scheduler_output)
else:
# For text-only models, we use token ids as input.
# While it is possible to use embeddings as input just like the
@@ -1537,7 +1532,7 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
# then the embedding layer is not included in the CUDA graph.
input_ids = self.input_ids[:num_input_tokens]
inputs_embeds = None
model_kwargs = {}
model_mm_kwargs = {}
if self.uses_mrope:
positions = self.mrope_positions[:, :num_input_tokens]
else:
@@ -1571,7 +1566,7 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
intermediate_tensors=intermediate_tensors,
inputs_embeds=inputs_embeds,
**MultiModalKwargs.as_kwargs(
model_kwargs,
model_mm_kwargs,
device=self.device,
),
)
@@ -2149,6 +2144,30 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
yield
input_ids.fill_(0)
def _get_mm_dummy_batch(
self,
modality: str,
max_items_per_batch: int,
) -> BatchedTensorInputs:
"""Dummy data for profiling and precompiling multimodal models."""
dummy_decoder_data = self.mm_registry.get_decoder_dummy_data(
model_config=self.model_config,
seq_len=self.max_num_tokens,
mm_counts={modality: 1},
)
dummy_mm_data = dummy_decoder_data.multi_modal_data
# Result in the maximum GPU consumption of the model
dummy_mm_item = dummy_mm_data.get_item(modality=modality, item_index=0)
dummy_mm_kwargs = MultiModalKwargs.from_items([dummy_mm_item])
batched_dummy_mm_inputs = MultiModalKwargs.batch([dummy_mm_kwargs] *
max_items_per_batch)
return MultiModalKwargs.as_kwargs(
batched_dummy_mm_inputs,
device=self.device,
)
@torch.inference_mode()
def _dummy_run(
self,
@@ -2213,16 +2232,14 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
with self.maybe_dummy_run_with_lora(self.lora_config,
num_scheduled_tokens):
model = self.model
if self.is_multimodal_model:
model_kwargs = self._init_model_kwargs_for_multimodal_model(
num_reqs=num_reqs)
input_ids = None
inputs_embeds = self.inputs_embeds[:num_tokens]
model_mm_kwargs = self._dummy_mm_kwargs(num_reqs)
else:
input_ids = self.input_ids[:num_tokens]
inputs_embeds = None
model_kwargs = {}
model_mm_kwargs = {}
if self.uses_mrope:
positions = self.mrope_positions[:, :num_tokens]
@@ -2247,13 +2264,13 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
self.vllm_config,
num_tokens=num_tokens,
num_tokens_across_dp=num_tokens_across_dp):
outputs = model(
outputs = self.model(
input_ids=input_ids,
positions=positions,
intermediate_tensors=intermediate_tensors,
inputs_embeds=inputs_embeds,
**MultiModalKwargs.as_kwargs(
model_kwargs,
model_mm_kwargs,
device=self.device,
),
)
@@ -2423,75 +2440,51 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
def profile_run(self) -> None:
# Profile with multimodal encoder & encoder cache.
# TODO: handle encoder-decoder models once we support them.
if (self.is_multimodal_model and self.max_num_encoder_input_tokens > 0
and self.encoder_cache_size > 0):
if self.is_multimodal_model:
mm_budget = self.mm_budget
assert mm_budget is not None
# NOTE: Currently model is profiled with a single non-text
# modality with the max possible input tokens even when
# it supports multiple.
max_tokens_by_modality_dict = self.mm_registry \
.get_max_tokens_per_item_by_nonzero_modality(self.model_config)
dummy_data_modality, max_tokens_per_mm_item = max(
max_tokens_by_modality_dict.items(), key=lambda item: item[1])
# TODO: handle encoder-decoder models once we support them.
if (encoder_budget := mm_budget.get_encoder_budget()) > 0:
# NOTE: Currently model is profiled with a single non-text
# modality with the max possible input tokens even when
# it supports multiple.
(
dummy_modality,
max_tokens,
) = mm_budget.get_modality_with_max_tokens()
(
max_mm_items_per_prompt,
max_mm_items_per_batch,
) = mm_budget.get_max_items(dummy_modality, max_tokens)
# Check how many items of this modality can be supported by
# the encoder budget.
encoder_budget = min(self.max_num_encoder_input_tokens,
self.encoder_cache_size)
logger.info(
"Encoder cache will be initialized with a budget of "
"%s tokens, and profiled with %s %s items of the maximum "
"feature size.",
encoder_budget,
max_mm_items_per_batch,
dummy_modality,
)
max_num_mm_items_encoder_budget = encoder_budget // \
max_tokens_per_mm_item
# Create dummy batch of multimodal inputs.
batched_dummy_mm_inputs = self._get_mm_dummy_batch(
dummy_modality,
max_mm_items_per_batch,
)
# Check how many items of this modality can be supported by
# the decoder budget.
max_mm_items_per_req = self.mm_registry.get_mm_limits_per_prompt(
self.model_config)[dummy_data_modality]
# Run multimodal encoder.
dummy_encoder_outputs = self.model.get_multimodal_embeddings(
**batched_dummy_mm_inputs)
# NOTE: We do not consider max_num_batched_tokens on purpose
# because the multimodal embeddings can be generated in advance
# and chunked prefilled.
max_num_mm_items_decoder_budget = self.max_num_reqs * \
max_mm_items_per_req
sanity_check_mm_encoder_outputs(
dummy_encoder_outputs,
expected_num_items=max_mm_items_per_batch,
)
max_num_mm_items = max(
1,
min(max_num_mm_items_encoder_budget,
max_num_mm_items_decoder_budget))
logger.info(
"Encoder cache will be initialized with a budget of %s tokens,"
" and profiled with %s %s items of the maximum feature size.",
encoder_budget, max_num_mm_items, dummy_data_modality)
# Create dummy batch of multimodal inputs.
dummy_mm_kwargs = self.mm_registry.get_decoder_dummy_data(
model_config=self.model_config,
seq_len=max_tokens_per_mm_item,
mm_counts={
dummy_data_modality: 1
},
).multi_modal_data
batched_dummy_mm_inputs = MultiModalKwargs.batch(
[dummy_mm_kwargs] * max_num_mm_items,
pin_memory=self.pin_memory)
batched_dummy_mm_inputs = MultiModalKwargs.as_kwargs(
batched_dummy_mm_inputs,
device=self.device,
)
# Run multimodal encoder.
dummy_encoder_outputs = self.model.get_multimodal_embeddings(
**batched_dummy_mm_inputs)
sanity_check_mm_encoder_outputs(
dummy_encoder_outputs,
expected_num_items=max_num_mm_items,
)
# Cache the dummy encoder outputs.
self.encoder_cache["tmp"] = dict(enumerate(dummy_encoder_outputs))
# Cache the dummy encoder outputs.
self.encoder_cache["tmp"] = dict(
enumerate(dummy_encoder_outputs))
# Add `is_profile` here to pre-allocate communication buffers
hidden_states, last_hidden_states \