vllm-project/vllm

[Transformers v5] MiniCPMV cannot apply processor

Open

#38,385 创建于 2026年3月27日

在 GitHub 查看
 (8 评论) (0 反应) (1 负责人)Python (80,034 star) (16,816 fork)batch import
good first issuehelp wanted

描述

This is a sub-issue forming part of the work in https://github.com/vllm-project/vllm/issues/38379, please read the description of this issue before beginning to work on this one.

Which test is failing?

If the model is relevant enough for us to keep supporting, this should be upstreamed to Transformers.

$ pytest tests/lora/test_minicpmv_tp.py::test_minicpmv_lora
...
ValueError: Failed to apply MiniCPMVProcessor on data={'text': '(<image>./</image>)', 'images': [<PIL.Image.Image image mode=RGB size=448x4032 at 0x7F7695579D60>]} with kwargs={'truncation': False, 'return_tensors': 'pt'}

How to configure my environment?

It's very important that you install both vLLM and Transformers from source so that your test results reflect the current state of both libraries.

# Or your fork
git clone https://github.com/huggingface/transformers.git
git clone https://github.com/vllm-project/vllm.git

cd vllm
VLLM_USE_PRECOMPILED=1 uv pip install -e .
uv pip install -e ../transformers

贡献者指南

技术栈
pythonpytorch
领域
backendmachine learning
议题类型
bug
难度面向新贡献者的预计实现难度，1 表示很小改动，5 表示专家级工作。
3
预计时间有经验贡献者完成调查、实现、测试并准备 pull request 的粗略时间范围。
half day
活动状态议题当前的可参与程度：新鲜、活跃、陈旧、阻塞或等待维护者输入。
active
清晰度议题是否清楚说明期望改动、验收标准和下一步。
mostly clear
前置要求
Pythontesting experience
新手友好度1-100 的估计分数，表示该议题对首次贡献者的友好程度。
40
研究方向
Investigate why MiniCPMVProcessor.apply fails on the given input. Look at the test file tests/lora/test minicpmv tp.py and the processor implementation in the transformers library. Check if the issue is related to argument handling in the processor, possibly the 'truncation' or 'return tensors' kwargs. Since this is part of a larger migration, refer to the parent issue vllm project/vllm#38379 for context. The fix may involve updating the processor call or adjusting the test inputs to match the expected format.