Closed
Description
Your current environment
The output of python collect_env.py
==============================
vLLM Info
==============================
ROCM Version : Could not collect
Neuron SDK Version : N/A
vLLM Version : 0.9.1.dev241+g8335667c2.d20250610 (git sha: 8335667c2, date: 20250610)
vLLM Build Flags:
CUDA Archs: Not Set; ROCm: Disabled; Neuron: Disabled
GPU Topology:
Could not collect
==============================
Environment Variables
==============================
NCCL_CUMEM_ENABLE=0
PYTORCH_NVML_BASED_CUDA_CHECK=1
TORCHINDUCTOR_COMPILE_THREADS=1
TORCHINDUCTOR_CACHE_DIR=/tmp/torchinductor_root
🐛 Describe the bug
When building the vllm-cpu-dev image using the command:
docker build -f docker/Dockerfile.cpu --target vllm-dev -t vllm-cpu-dev . the build fails with the following error:
0.193 Using Python 3.12.11 environment at: /opt/venv
9.622 × No solution found when resolving dependencies:
9.622 ╰─▶ Because there is no version of torch==2.7.0+cu128 and you require
9.622 torch==2.7.0+cu128, we can conclude that your requirements are
9.622 unsatisfiable.
The error occurs during the uv pip install -r requirements/dev.txt step.
Steps to Reproduce:
- Clone the vllm repository.
- Run the command:
docker build -f docker/Dockerfile.cpu --target vllm-dev -t vllm-cpu-dev .
Proposed Fix:
Modify Dockerfile.cpu, uv compile test.in before uv pip install all the test dependency.
Before submitting a new issue...
- Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.