A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 12, 07:15 UTC
[CI] Skip paddleocr_vl for transformer 4.57.3 (#29758)
Signed-off-by: Huamin Li <3ericli@gmail.com>
Huamin Li committed
Nov 30, 2025 at 20:38 UTC
83805a6078c73f7f72d270952965a637c40ebdf2
1 file changed
+6
tests/models/multimodal/generation/test_common.py
+6
@@ -708,6 +708,12 @@ VLM_TEST_SETTINGS = {
708
max_num_seqs=2,
709
auto_cls=AutoModelForCausalLM,
710
image_size_factors=[(), (0.25,)],
711
+ marks=[
712
+ pytest.mark.skipif(
713
+ Version(TRANSFORMERS_VERSION) == Version("4.57.3"),
714
+ reason="This model is broken in Transformers v4.57.3",
715
+ )
716
+ ],
717
),
718
"phi3v": VLMTestInfo(
719
models=["microsoft/Phi-3.5-vision-instruct"],