[MM] Remove text components from ProcessorInputs (#53093)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Cyrus Leung
2026-08-20 20:42:53 -07:00
committed by GitHub
parent 2adf4b9e2a
commit b00f475f09
92 changed files with 220 additions and 806 deletions
@@ -11,6 +11,7 @@ from pqdm.threads import pqdm
from vllm.multimodal import MULTIMODAL_REGISTRY
from vllm.multimodal.parse import ImageSize
from vllm.multimodal.processing import BaseMultiModalProcessor
from vllm.tokenizers.hf import maybe_make_thread_pool
from ...utils import build_model_context
@@ -144,7 +145,14 @@ def test_processor_prompt_replacements_regression(model_id, num_imgs):
mm_processor_kwargs=None,
limit_mm_per_prompt={"image": num_imgs},
)
processor = MULTIMODAL_REGISTRY.create_processor(ctx.model_config)
# Avoid tokenizer already borrowed error
maybe_make_thread_pool(ctx.tokenizer)
processor = MULTIMODAL_REGISTRY.create_processor(
ctx.model_config,
tokenizer=ctx.tokenizer,
)
image_ratios = [
(171, 152),
@@ -177,7 +185,14 @@ def test_processor_prompt_replacements_all(model_id, num_imgs):
mm_processor_kwargs=None,
limit_mm_per_prompt={"image": num_imgs},
)
processor = MULTIMODAL_REGISTRY.create_processor(ctx.model_config)
# Avoid tokenizer already borrowed error
maybe_make_thread_pool(ctx.tokenizer)
processor = MULTIMODAL_REGISTRY.create_processor(
ctx.model_config,
tokenizer=ctx.tokenizer,
)
seen_aspect_ratios = set[float]()
image_sizes = list[ImageSize]()