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

Fix encoder-decoder model disabling mm processor cache (#33236)

Browse Source 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Harry Mellor
2026-01-30 16:30:10 +00:00
committed by GitHub
parent 8ece60768f
commit 67239c4c42
2 changed files with 4 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
tests/models/multimodal/processing/test_common.py
View File

@@ -241,14 +241,15 @@ def _test_processing_correctness(
revision=model_info.revision,
trust_remote_code=model_info.trust_remote_code,
hf_overrides=model_info.hf_overrides,
# Ensure that the cache can fit all of the data
mm_processor_cache_gb=2048,
skip_tokenizer_init=model_info.require_embed_inputs,
enable_prompt_embeds=model_info.require_embed_inputs,
enable_mm_embeds=model_info.require_embed_inputs,
enforce_eager=model_info.enforce_eager,
dtype=model_info.dtype,
)
# Ensure that the cache can fit all of the data
# (set after because ModelConfig would set it to 0 for encoder-decoder models)
model_config.multimodal_config.mm_processor_cache_gb = 2048
model_cls = MULTIMODAL_REGISTRY._get_model_cls(model_config)
factories = model_cls._processor_factory