Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[V1][PP] Cache Intermediate Tensors #13353

Merged
merged 2 commits into from
Feb 16, 2025
+28 −8
Merged

[V1][PP] Cache Intermediate Tensors #13353

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
e188540
[V1][PP] Cache Intermediate Tensors
WoosukKwon Feb 16, 2025
766541d
minor
WoosukKwon Feb 16, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
36 changes: 28 additions & 8 deletions vllm/v1/worker/gpu_model_runner.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@

import gc
import time
from typing import TYPE_CHECKING, Dict, List, Optional, Tuple, cast
from typing import TYPE_CHECKING, Dict, List, Optional, Tuple, Union, cast

import numpy as np
import torch
Expand Down Expand Up @@ -149,6 +149,7 @@ def __init__(
self.positions = torch.zeros(self.max_num_tokens,
dtype=torch.int64,
device=self.device)
# self.intermediate_tensors # Set after load_model

# Only relevant for models using M-RoPE (e.g, Qwen2-VL)
if self.uses_mrope:
Expand Down Expand Up @@ -869,7 +870,7 @@ def execute_model(
self,
scheduler_output: "SchedulerOutput",
intermediate_tensors: Optional[IntermediateTensors] = None,
) -> ModelRunnerOutput:
) -> Union[ModelRunnerOutput, torch.Tensor]:
batch_changed = self._update_states(scheduler_output)

if self.is_multimodal_model:
Expand Down Expand Up @@ -919,6 +920,14 @@ def execute_model(
else:
positions = self.positions[:num_input_tokens]

if get_pp_group().is_first_rank:
intermediate_tensors = None
else:
intermediate_tensors = IntermediateTensors({
k: v[:num_input_tokens]
for k, v in self.intermediate_tensors.items()
})

# Run the decoder.
# Use persistent buffers for CUDA graphs.
with set_forward_context(attn_metadata, self.vllm_config):
Expand All @@ -931,7 +940,9 @@ def execute_model(
inputs_embeds=inputs_embeds,
)
if not get_pp_group().is_last_rank:
# For mid-pipeline stages, return the hidden states.
return hidden_states

hidden_states = hidden_states[:num_scheduled_tokens]
sample_hidden_states = hidden_states[logits_indices]
logits = self.model.compute_logits(sample_hidden_states, None)
Expand Down Expand Up @@ -1118,12 +1129,21 @@ def _dummy_run(
positions = self.mrope_positions[:, :num_tokens]
else:
positions = self.positions[:num_tokens]
intermediate_tensors = None
if not get_pp_group().is_first_rank:
intermediate_tensors = self.model.make_empty_intermediate_tensors(
batch_size=num_tokens,
dtype=self.model_config.dtype,
device=self.device)

if get_pp_group().is_first_rank:
intermediate_tensors = None
else:
if not hasattr(self, "intermediate_tensors"):
self.intermediate_tensors = (
self.model.make_empty_intermediate_tensors(
batch_size=self.max_num_tokens,
dtype=self.model_config.dtype,
device=self.device))
intermediate_tensors = IntermediateTensors({
k: v[:num_tokens]
for k, v in self.intermediate_tensors.items()
})

with set_forward_context(None, self.vllm_config):
hidden_states = model(
input_ids=input_ids,
Expand Down