biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[VLM][Bugfix] Pass processor kwargs properly on init (#13516)

Browse Source 
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Cyrus Leung
2025-02-19 21:13:50 +08:00
committed by GitHub
parent 52ce14d31f
commit 377d10bd14
44 changed files with 677 additions and 455 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
vllm/model_executor/models/paligemma.py
View File

@@ -16,8 +16,8 @@ from vllm.model_executor.layers.sampler import SamplerOutput
from vllm.model_executor.sampling_metadata import SamplingMetadata
from vllm.multimodal import MULTIMODAL_REGISTRY
from vllm.multimodal.inputs import NestedTensors
from vllm.multimodal.utils import cached_get_tokenizer
from vllm.sequence import IntermediateTensors
from vllm.transformers_utils.tokenizer import cached_tokenizer_from_config
from .interfaces import SupportsMultiModal, SupportsPP
from .siglip import (SiglipVisionModel, dummy_image_for_siglip,
@@ -88,7 +88,7 @@ def input_processor_for_paligemma(ctx: InputContext,
model_config = ctx.model_config
hf_config = ctx.get_hf_config(PaliGemmaConfig)
tokenizer = cached_get_tokenizer(model_config.tokenizer)
tokenizer = cached_tokenizer_from_config(model_config)
image_feature_size = hf_config.text_config.num_image_tokens
image_token_str = tokenizer.decode(hf_config.image_token_index)
bos_token = tokenizer.decode(hf_config.bos_token_id)