Skip to content

[Multi Modal] Update multimodal_manager to align with vllm changes #1618

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

Merged
kwang3939 merged 1 commit into from
Feb 4, 2026
Merged

[Multi Modal] Update multimodal_manager to align with vllm changes #1618

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
2 changes: 1 addition & 1 deletion .buildkite/vllm_lkg.version
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1 +1 @@
64a40a7ab4d0053830fae04c83763fa67f2183e6
010ec0c30ef1fa481fc69b8a5d2c205052d93607
22 changes: 11 additions & 11 deletions tpu_inference/runner/multimodal_manager.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,8 +20,6 @@
from vllm.multimodal.inputs import MultiModalKwargsItem, PlaceholderRange
from vllm.multimodal.utils import group_mm_kwargs_by_modality
from vllm.v1.core.sched.output import SchedulerOutput as VllmSchedulerOutput
from vllm.v1.worker.utils import (gather_mm_placeholders,
scatter_mm_placeholders)

from tpu_inference.models.jax.utils.multi_modal_utils import (
flatten_embeddings, sanity_check_mm_encoder_outputs)
Expand Down Expand Up @@ -160,15 +158,12 @@ def execute_mm_encoder(self, scheduler_output: "VllmSchedulerOutput"):
encoder_outputs.append(output)

# Cache the encoder outputs.
for (mm_hash, pos_info), output in zip(
for (mm_hash, _), output in zip(
mm_hashes_pos,
encoder_outputs,
):

self.runner.encoder_cache[mm_hash] = scatter_mm_placeholders(
output,
is_embed=pos_info.is_embed,
)
self.runner.encoder_cache[mm_hash] = output

def gather_mm_embeddings(self, scheduler_output: "VllmSchedulerOutput",
target_pad_len: int) -> list[jax.Array]:
Expand Down Expand Up @@ -201,6 +196,11 @@ def gather_mm_embeddings(self, scheduler_output: "VllmSchedulerOutput",
num_computed_tokens - start_pos + num_scheduled_tokens,
num_encoder_tokens)
assert start_idx < end_idx
curr_embeds_start, curr_embeds_end = (
pos_info.get_embeds_indices_in_range(start_idx, end_idx))
if curr_embeds_start == curr_embeds_end:
continue

mm_hash = mm_feature.identifier
encoder_output = self.runner.encoder_cache.get(mm_hash, None)
assert encoder_output is not None,\
Expand All @@ -209,11 +209,11 @@ def gather_mm_embeddings(self, scheduler_output: "VllmSchedulerOutput",

if (is_embed := pos_info.is_embed) is not None:
is_embed = is_embed[start_idx:end_idx]
mm_embeds_item = encoder_output[
curr_embeds_start:curr_embeds_end]
else:
mm_embeds_item = encoder_output[start_idx:end_idx]

mm_embeds_item = gather_mm_placeholders(
encoder_output[start_idx:end_idx],
is_embed=is_embed,
)
mm_embeds.append(mm_embeds_item)
if not mm_embeds:
return None
Expand Down
Loading