[Misc][LLaMa4] Compile LLaMa Vision Encoder (#30709)
Signed-off-by: Lucas Kabela <lucaskabela@meta.com>
This commit is contained in:
@@ -71,3 +71,40 @@ def test_qwen2_5_vl_no_vit_compilation(vllm_runner, monkeypatch):
|
|||||||
) as _,
|
) as _,
|
||||||
):
|
):
|
||||||
pass
|
pass
|
||||||
|
|
||||||
|
|
||||||
|
# forked needed to workaround https://github.com/vllm-project/vllm/issues/21073
|
||||||
|
# Requires Cuda and 8 gpus as well
|
||||||
|
@pytest.mark.forked
|
||||||
|
@pytest.mark.skip(reason="Skipping due to CI resource constraints")
|
||||||
|
def test_mllama4_vit_compilation(vllm_runner, monkeypatch):
|
||||||
|
"""Test that Mllama4 vision submodules are compiled.
|
||||||
|
|
||||||
|
This test verifies that the 2 vision submodules (Llama4VisionEncoder,
|
||||||
|
Llama4VisionPixelShuffleMLP) are properly tagged
|
||||||
|
for compilation by checking that num_models_seen increases to 3.
|
||||||
|
|
||||||
|
However since we are using TP=8, we compilation_counter will not
|
||||||
|
work properly so we will just check the run succeeds rn
|
||||||
|
"""
|
||||||
|
# Disable multiprocessing so that the counter is in the same process
|
||||||
|
monkeypatch.setenv("VLLM_ENABLE_V1_MULTIPROCESSING", "0")
|
||||||
|
|
||||||
|
with (
|
||||||
|
monkeypatch.context(),
|
||||||
|
# TODO: Since we require TP=8, this messes with the compilation
|
||||||
|
# counter. We should fix this in the future, but leave for now
|
||||||
|
# to make sure that compilation runs (no crash) with llama vision encoder
|
||||||
|
compilation_counter.expect(num_models_seen=0),
|
||||||
|
vllm_runner(
|
||||||
|
"meta-llama/Llama-4-Scout-17B-16E-Instruct",
|
||||||
|
max_model_len=512,
|
||||||
|
gpu_memory_utilization=0.8,
|
||||||
|
tensor_parallel_size=8,
|
||||||
|
compilation_config={
|
||||||
|
"mode": CompilationMode.VLLM_COMPILE,
|
||||||
|
"compile_mm_encoder": True,
|
||||||
|
},
|
||||||
|
),
|
||||||
|
):
|
||||||
|
pass
|
||||||
|
|||||||
@@ -430,8 +430,9 @@ class CompilationConfig:
|
|||||||
If empty list [], no ops are excluded (suitable for full cudagraphs)."""
|
If empty list [], no ops are excluded (suitable for full cudagraphs)."""
|
||||||
compile_mm_encoder: bool = False
|
compile_mm_encoder: bool = False
|
||||||
"""Whether or not to compile the multimodal encoder.
|
"""Whether or not to compile the multimodal encoder.
|
||||||
Currently, this only works for `Qwen2_5_vl` on selected platforms.
|
Currently, this only works for `Qwen2_5_vl` and `mLLaMa4` models
|
||||||
Disabled by default until more models are supported/tested to work."""
|
on selected platforms. Disabled by default until more models
|
||||||
|
are supported/tested to work."""
|
||||||
|
|
||||||
# Inductor capture
|
# Inductor capture
|
||||||
compile_sizes: list[int | str] | None = None
|
compile_sizes: list[int | str] | None = None
|
||||||
|
|||||||
@@ -171,12 +171,12 @@ class MMEncoderAttention(CustomOp):
|
|||||||
q=query,
|
q=query,
|
||||||
k=key,
|
k=key,
|
||||||
v=value,
|
v=value,
|
||||||
scale=self.scale,
|
|
||||||
cu_seqlens=cu_seqlens,
|
|
||||||
max_seqlen=max_seqlen,
|
|
||||||
batch_size=bsz,
|
batch_size=bsz,
|
||||||
is_rocm_aiter=(self.attn_backend == AttentionBackendEnum.ROCM_AITER_FA),
|
is_rocm_aiter=(self.attn_backend == AttentionBackendEnum.ROCM_AITER_FA),
|
||||||
fa_version=self._fa_version,
|
fa_version=self._fa_version,
|
||||||
|
scale=self.scale,
|
||||||
|
cu_seqlens=cu_seqlens,
|
||||||
|
max_seqlen=max_seqlen,
|
||||||
)
|
)
|
||||||
if is_reshaped:
|
if is_reshaped:
|
||||||
output = output.reshape(bsz, q_len, -1)
|
output = output.reshape(bsz, q_len, -1)
|
||||||
|
|||||||
@@ -60,14 +60,17 @@ class Llama4VisionRotaryEmbedding(RotaryEmbeddingBase):
|
|||||||
assert key is not None
|
assert key is not None
|
||||||
# self.cos_sin_cache here is complex tensor so we cannot cast into
|
# self.cos_sin_cache here is complex tensor so we cannot cast into
|
||||||
# query's dtype directly with self._match_cos_sin_cache_dtype
|
# query's dtype directly with self._match_cos_sin_cache_dtype
|
||||||
self.cos_sin_cache: torch.Tensor = self.cos_sin_cache.to(query.device)
|
|
||||||
|
# NOTE: by not storing cos_sin_cache in self, we can avoid
|
||||||
|
# memory buffer update which is costly to runtime
|
||||||
|
cos_sin_cache: torch.Tensor = self.cos_sin_cache.to(query.device)
|
||||||
query_ = torch.view_as_complex(query.float().reshape(*query.shape[:-1], -1, 2))
|
query_ = torch.view_as_complex(query.float().reshape(*query.shape[:-1], -1, 2))
|
||||||
key_ = torch.view_as_complex(key.float().reshape(*key.shape[:-1], -1, 2))
|
key_ = torch.view_as_complex(key.float().reshape(*key.shape[:-1], -1, 2))
|
||||||
broadcast_shape = [
|
broadcast_shape = [
|
||||||
d if i == 1 or i == (query_.ndim - 1) else 1
|
d if i == 1 or i == (query_.ndim - 1) else 1
|
||||||
for i, d in enumerate(query_.shape)
|
for i, d in enumerate(query_.shape)
|
||||||
]
|
]
|
||||||
freqs_ci = self.cos_sin_cache.view(*broadcast_shape)
|
freqs_ci = cos_sin_cache.view(*broadcast_shape)
|
||||||
query_out = torch.view_as_real(query_ * freqs_ci).flatten(3)
|
query_out = torch.view_as_real(query_ * freqs_ci).flatten(3)
|
||||||
key_out = torch.view_as_real(key_ * freqs_ci).flatten(3)
|
key_out = torch.view_as_real(key_ * freqs_ci).flatten(3)
|
||||||
return query_out.type_as(query), key_out.type_as(key)
|
return query_out.type_as(query), key_out.type_as(key)
|
||||||
|
|||||||
@@ -369,7 +369,11 @@ def llama_model_invariants(
|
|||||||
torch._check(positions.size()[0] == input_ids.size()[0])
|
torch._check(positions.size()[0] == input_ids.size()[0])
|
||||||
|
|
||||||
|
|
||||||
@support_torch_compile(shape_invariants=llama_model_invariants)
|
@support_torch_compile(
|
||||||
|
# TODO[#32068]: Investigate recompilation
|
||||||
|
# mark_unbacked_dims={"input_ids": 0},
|
||||||
|
shape_invariants=llama_model_invariants
|
||||||
|
)
|
||||||
class LlamaModel(nn.Module):
|
class LlamaModel(nn.Module):
|
||||||
def __init__(
|
def __init__(
|
||||||
self,
|
self,
|
||||||
|
|||||||
@@ -31,9 +31,11 @@ from transformers.models.llama4.image_processing_llama4_fast import (
|
|||||||
get_best_fit,
|
get_best_fit,
|
||||||
)
|
)
|
||||||
|
|
||||||
from vllm.config import VllmConfig
|
from vllm.compilation.decorators import support_torch_compile
|
||||||
|
from vllm.config import VllmConfig, set_current_vllm_config
|
||||||
from vllm.config.multimodal import BaseDummyOptions
|
from vllm.config.multimodal import BaseDummyOptions
|
||||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||||
|
from vllm.forward_context import set_forward_context
|
||||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||||
from vllm.model_executor.layers.fused_moe import FusedMoE
|
from vllm.model_executor.layers.fused_moe import FusedMoE
|
||||||
from vllm.model_executor.layers.linear import (
|
from vllm.model_executor.layers.linear import (
|
||||||
@@ -47,6 +49,7 @@ from vllm.model_executor.layers.rotary_embedding import get_rope
|
|||||||
from vllm.model_executor.model_loader.utils import initialize_model
|
from vllm.model_executor.model_loader.utils import initialize_model
|
||||||
from vllm.model_executor.model_loader.weight_utils import default_weight_loader
|
from vllm.model_executor.model_loader.weight_utils import default_weight_loader
|
||||||
from vllm.model_executor.models.module_mapping import MultiModelKeys
|
from vllm.model_executor.models.module_mapping import MultiModelKeys
|
||||||
|
from vllm.model_executor.models.vision import should_torch_compile_mm_vit
|
||||||
from vllm.multimodal import MULTIMODAL_REGISTRY
|
from vllm.multimodal import MULTIMODAL_REGISTRY
|
||||||
from vllm.multimodal.inputs import (
|
from vllm.multimodal.inputs import (
|
||||||
MultiModalDataDict,
|
MultiModalDataDict,
|
||||||
@@ -456,6 +459,9 @@ class Llama4UnfoldConvolution(nn.Module):
|
|||||||
return hidden_states
|
return hidden_states
|
||||||
|
|
||||||
|
|
||||||
|
@support_torch_compile(
|
||||||
|
dynamic_arg_dims={"images_flattened": 0}, enable_if=should_torch_compile_mm_vit
|
||||||
|
)
|
||||||
class Llama4VisionModel(nn.Module):
|
class Llama4VisionModel(nn.Module):
|
||||||
def __init__(
|
def __init__(
|
||||||
self,
|
self,
|
||||||
@@ -497,6 +503,7 @@ class Llama4VisionModel(nn.Module):
|
|||||||
prefix=f"{prefix}.model",
|
prefix=f"{prefix}.model",
|
||||||
use_data_parallel=use_data_parallel,
|
use_data_parallel=use_data_parallel,
|
||||||
)
|
)
|
||||||
|
|
||||||
self.vision_adapter = Llama4VisionPixelShuffleMLP(
|
self.vision_adapter = Llama4VisionPixelShuffleMLP(
|
||||||
config,
|
config,
|
||||||
quant_config,
|
quant_config,
|
||||||
@@ -762,18 +769,28 @@ class Llama4ForConditionalGeneration(
|
|||||||
multimodal_config = vllm_config.model_config.multimodal_config
|
multimodal_config = vllm_config.model_config.multimodal_config
|
||||||
self.use_data_parallel = multimodal_config.mm_encoder_tp_mode == "data"
|
self.use_data_parallel = multimodal_config.mm_encoder_tp_mode == "data"
|
||||||
|
|
||||||
|
self.vllm_config = vllm_config
|
||||||
self.config = config
|
self.config = config
|
||||||
self.quant_config = quant_config
|
self.quant_config = quant_config
|
||||||
self.multimodal_config = multimodal_config
|
self.multimodal_config = multimodal_config
|
||||||
if multimodal_config.get_limit_per_prompt("image"):
|
if multimodal_config.get_limit_per_prompt("image"):
|
||||||
self.vision_model = Llama4VisionModel(
|
from vllm.compilation.backends import set_model_tag
|
||||||
config.vision_config,
|
|
||||||
None,
|
with (
|
||||||
prefix=maybe_prefix(prefix, "vision_model"),
|
set_current_vllm_config(vllm_config),
|
||||||
use_data_parallel=self.use_data_parallel,
|
set_model_tag("Llama4VisionModel", is_encoder=True),
|
||||||
)
|
):
|
||||||
|
self.vision_model = Llama4VisionModel(
|
||||||
|
config=config.vision_config,
|
||||||
|
quant_config=None,
|
||||||
|
prefix=maybe_prefix(prefix, "vision_model"),
|
||||||
|
use_data_parallel=self.use_data_parallel,
|
||||||
|
)
|
||||||
|
|
||||||
self.multi_modal_projector = Llama4MultiModalProjector(
|
self.multi_modal_projector = Llama4MultiModalProjector(
|
||||||
self.config, None, prefix=maybe_prefix(prefix, "multi_modal_projector")
|
config=self.config,
|
||||||
|
quant_config=None,
|
||||||
|
prefix=maybe_prefix(prefix, "multi_modal_projector"),
|
||||||
)
|
)
|
||||||
else:
|
else:
|
||||||
self.vision_model = None
|
self.vision_model = None
|
||||||
@@ -883,7 +900,10 @@ class Llama4ForConditionalGeneration(
|
|||||||
if image_input is None:
|
if image_input is None:
|
||||||
return []
|
return []
|
||||||
|
|
||||||
return self._process_image_input(image_input)
|
with (
|
||||||
|
set_forward_context(None, self.vllm_config),
|
||||||
|
):
|
||||||
|
return self._process_image_input(image_input)
|
||||||
|
|
||||||
def forward(
|
def forward(
|
||||||
self,
|
self,
|
||||||
|
|||||||
@@ -72,9 +72,9 @@ def flash_attn_maxseqlen_wrapper_fake(
|
|||||||
batch_size: int,
|
batch_size: int,
|
||||||
is_rocm_aiter: bool,
|
is_rocm_aiter: bool,
|
||||||
fa_version: int | None,
|
fa_version: int | None,
|
||||||
scale: float | None,
|
scale: float | None = None,
|
||||||
cu_seqlens: torch.Tensor | None,
|
cu_seqlens: torch.Tensor | None = None,
|
||||||
max_seqlen: torch.Tensor | None,
|
max_seqlen: torch.Tensor | None = None,
|
||||||
) -> torch.Tensor:
|
) -> torch.Tensor:
|
||||||
return torch.empty_like(q)
|
return torch.empty_like(q)
|
||||||
|
|
||||||
|
|||||||
Reference in New Issue
Block a user