[Installation]: vLLM nightly version(0.11.1) is unable to install via uv according to vLLM recipes for DeepSeek-OCR models

[KnightChaser]

Your current environment

Note: I ran the following script without Python3 virtual environment.

==============================
        System Info
==============================
OS                           : Debian GNU/Linux 12 (bookworm) (x86_64)
GCC version                  : (Debian 12.2.0-14+deb12u1) 12.2.0
Clang version                : 14.0.6
CMake version                : version 3.25.1
Libc version                 : glibc-2.36

==============================
       PyTorch Info
==============================
PyTorch version              : 2.8.0+cu128
Is debug build               : False
CUDA used to build PyTorch   : 12.8
ROCM used to build PyTorch   : N/A

==============================
      Python Environment
==============================
Python version               : 3.12.12 (main, Oct 31 2025, 23:02:31) [Clang 21.1.4 ] (64-bit runtime)
Python platform              : Linux-6.1.0-30-amd64-x86_64-with-glibc2.36

==============================
       CUDA / GPU Info
==============================
Is CUDA available            : True
CUDA runtime version         : 11.8.89
CUDA_MODULE_LOADING set to   : LAZY
GPU models and configuration :
GPU 0: NVIDIA RTX A6000
GPU 1: NVIDIA RTX A6000

Nvidia driver version        : 555.42.02
cuDNN version                : Could not collect
HIP runtime version          : N/A
MIOpen runtime version       : N/A
Is XNNPACK available         : True

==============================
          CPU Info
==============================
Architecture:                         x86_64
CPU op-mode(s):                       32-bit, 64-bit
Address sizes:                        45 bits physical, 48 bits virtual
Byte Order:                           Little Endian
CPU(s):                               16
On-line CPU(s) list:                  0-15
Vendor ID:                            GenuineIntel
Model name:                           Intel(R) Xeon(R) Gold 6326 CPU @ 2.90GHz
CPU family:                           6
Model:                                106
Thread(s) per core:                   1
Core(s) per socket:                   16
Socket(s):                            1
Stepping:                             6
BogoMIPS:                             4199.99
Flags:                                fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon rep_good nopl xtopology tsc_reliable nonstop_tsc cpuid tsc_known_freq pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch invpcid_single ssbd ibrs ibpb stibp ibrs_enhanced fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves wbnoinvd arat avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq rdpid fsrm md_clear flush_l1d arch_capabilities
Hypervisor vendor:                    VMware
Virtualization type:                  full
L1d cache:                            768 KiB (16 instances)
L1i cache:                            512 KiB (16 instances)
L2 cache:                             20 MiB (16 instances)
L3 cache:                             24 MiB (1 instance)
NUMA node(s):                         1
NUMA node0 CPU(s):                    0-15
Vulnerability Gather data sampling:   Unknown: Dependent on hypervisor status
Vulnerability Itlb multihit:          KVM: Mitigation: VMX unsupported
Vulnerability L1tf:                   Not affected
Vulnerability Mds:                    Not affected
Vulnerability Meltdown:               Not affected
Vulnerability Mmio stale data:        Vulnerable: Clear CPU buffers attempted, no microcode; SMT Host state unknown
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed:               Not affected
Vulnerability Spec rstack overflow:   Not affected
Vulnerability Spec store bypass:      Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:             Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:             Mitigation; Enhanced / Automatic IBRS; IBPB conditional; RSB filling; PBRSB-eIBRS SW sequence; BHI SW loop, KVM SW loop
Vulnerability Srbds:                  Not affected
Vulnerability Tsx async abort:        Not affected

==============================
Versions of relevant libraries
==============================
[pip3] numpy==2.2.6
[pip3] nvidia-cublas-cu12==12.8.4.1
[pip3] nvidia-cuda-cupti-cu12==12.8.90
[pip3] nvidia-cuda-nvrtc-cu12==12.8.93
[pip3] nvidia-cuda-runtime-cu12==12.8.90
[pip3] nvidia-cudnn-cu12==9.10.2.21
[pip3] nvidia-cufft-cu12==11.3.3.83
[pip3] nvidia-cufile-cu12==1.13.1.3
[pip3] nvidia-curand-cu12==10.3.9.90
[pip3] nvidia-cusolver-cu12==11.7.3.90
[pip3] nvidia-cusparse-cu12==12.5.8.93
[pip3] nvidia-cusparselt-cu12==0.7.1
[pip3] nvidia-nccl-cu12==2.27.3
[pip3] nvidia-nvjitlink-cu12==12.8.93
[pip3] nvidia-nvtx-cu12==12.8.90
[pip3] pyzmq==27.1.0
[pip3] torch==2.8.0
[pip3] torchaudio==2.8.0
[pip3] torchvision==0.23.0
[pip3] transformers==4.57.1
[pip3] triton==3.4.0
[conda] Could not collect

==============================
         vLLM Info
==============================
ROCM Version                 : Could not collect
vLLM Version                 : 0.11.0
vLLM Build Flags:
  CUDA Archs: Not Set; ROCm: Disabled
GPU Topology:
        GPU0    GPU1    CPU Affinity    NUMA Affinity   GPU NUMA ID
GPU0     X      PHB     0-15    0               N/A
GPU1    PHB      X      0-15    0               N/A

Legend:

  X    = Self
  SYS  = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI)
  NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node
  PHB  = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU)
  PXB  = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge)
  PIX  = Connection traversing at most a single PCIe bridge
  NV#  = Connection traversing a bonded set of # NVLinks

==============================
     Environment Variables
==============================
CUDA_VISIBLE_DEVICES=0,1
CUDA_VISIBLE_DEVICES=0,1
PYTORCH_NVML_BASED_CUDA_CHECK=1
TORCHINDUCTOR_COMPILE_THREADS=1
CUDA_MODULE_LOADING=LAZY

How you are installing vllm

According to the vLLM Recipes page to try out DeepSeek-OCR model, with uv as a Python3 package manager, I created a virtual environment and tried to install the Nightly version (targeting 0.11.1, but currently the latest release version of vLLM is 0.11.0.) of the vLLM library that includes vllm/model_executor/models/deepseek_ocr.py.

uv venv
source .venv/bin/activate
uv pip install -U vllm --pre --extra-index-url https://wheels.vllm.ai/nightly

However, when I actually try that uv pip install ... command, I get the following installation failure error message.

(uvvenv) ____@____:~/ghrepo/pdfscribe2ds$ uv pip install -U vllm --pre --extra-index-url https://wheels.vllm.ai/nightly
  × No solution found when resolving dependencies:
  ╰─▶ Because there is no version of xformers{platform_machine == 'x86_64' and sys_platform == 'linux'}==0.0.33+5d4b92a5.d20251029
      and vllm==0.11.1rc6.dev85+g53f6e81df.cu129 depends on xformers{platform_machine == 'x86_64' and sys_platform ==
      'linux'}==0.0.33+5d4b92a5.d20251029, we can conclude that vllm==0.11.1rc6.dev85+g53f6e81df.cu129 cannot be used.
      And because only vllm==0.11.1rc6.dev85+g53f6e81df.cu129 is available and you require vllm, we can conclude that your requirements
      are unsatisfiable.

      hint: `vllm` was found on https://wheels.vllm.ai/nightly, but not at the requested version (all of:
          vllm<0.11.1rc6.dev85+g53f6e81df.cu129
          vllm>0.11.1rc6.dev85+g53f6e81df.cu129
      ). A compatible version may be available on a subsequent index (e.g., https://pypi.org/simple). By default, uv will only consider
      versions that are published on the first index that contains a given package, to avoid dependency confusion attacks. If all indexes
      are equally trusted, use `--index-strategy unsafe-best-match` to consider all versions from all indexes, regardless of the order in
      which they were defined.

If I just go with it with --index-strategy unsafe-best-match even though it's not technically desired,

uv pip install -U vllm --pre --extra-index-url https://wheels.vllm.ai/nightly --index-strategy unsafe-best-match

Then I get the following error that the DeepSeek-related library component doesn't exist. (I suppose that 0.11.0 version was installed instead of 0.11.1.)

INFO 11-04 17:16:02 [__init__.py:216] Automatically detected platform cuda.
Traceback (most recent call last):
  File "/home/grlee/ghrepo/pdfscribe2ds/app.py", line 8, in <module>
    from ocr_pipeline.pipeline import run_pdf_pipeline
  File "/home/grlee/ghrepo/pdfscribe2ds/ocr_pipeline/pipeline.py", line 12, in <module>
    from .ocr_engine import DeepSeekOCREngine
  File "/home/grlee/ghrepo/pdfscribe2ds/ocr_pipeline/ocr_engine.py", line 8, in <module>
    from vllm.model_executor.models.deepseek_ocr import NGramPerReqLogitsProcessor
ModuleNotFoundError: No module named 'vllm.model_executor.models.deepseek_ocr'

It looks like multiple people on the related Reddit forum is reporting the similar issues: https://www.reddit.com/r/Vllm/comments/1ol1njm/vllm_deepseekocr/

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.

[ZJY0516]

CC @ywang96

[okozarva]

Hi did you find any workaround?

[bash99]

I can install it with

uv pip install -U vllm --pre --extra-index-url https://wheels.vllm.ai/nightly "triton-kernels@git+https://github.com/triton-lang/triton.git@v3.5.0#subdirectory=python/triton_kernels"

uv only complains that triton depends need be set as above, which I obeys.

[blackboxradiology]

I had luck with this

pip install -U vllm --pre --extra-index-url https://wheels.vllm.ai/nightly --extra-index-url https://download.pytorch.org/whl/cu129

[KnightChaser]

I had luck with this

pip install -U vllm --pre --extra-index-url https://wheels.vllm.ai/nightly --extra-index-url https://download.pytorch.org/whl/cu129

Thanks @blackboxradiology, I also had the luck with you. When someone is trying to use that command with uv, the Python 3 package manager, then proceed with --index-strategy unsafe-best-match, and it will still work. (At least in my case)

uv pip install -U vllm --pre --extra-index-url https://wheels.vllm.ai/nightly --extra-index-url https://download.pytorch.org/whl/cu129 --index-strategy unsafe-best-match

After installing vLLM with that command, I could verify myself that the desired version of vLLM(0.11.1) has been installed.

... $ uv pip freeze | grep vllm
vllm==0.11.1rc6.dev107+g878fd5a16.cu129

Well, still, since it's not officially released version, that version may contain some unexpected problems, so I wish the official release of vLLM v0.11.1 quickly comes out!

[yihengforyishu]

same with this

[liyucheng09]

where can we find a 12.8 pre-built nightly wheel??

[ywang96]

where can we find a 12.8 pre-built nightly wheel??

@liyucheng09 We ran into some issue with xformers dependency, but the following command should work for 12.8

uv pip install -U vllm --pre --extra-index-url https://wheels.vllm.ai/nightly --extra-index-url https://download.pytorch.org/whl/cu128 --index-strategy unsafe-best-match

[liyucheng09]

@ywang96 thanks for the prompt response, but seems it was using cu129

(DeepSeek-OCR) /cpfs02/user/lyc508653/codespace/DeepSeek-OCR> uv pip install -U vllm --pre --extra-index-url https://wheels.vllm.ai/nightly --extra-index-url https://download.pytorch.org/whl/cu128
⠴ vllm==0.11.1rc6.dev183+g9da9208b2.cu129   

[liyucheng09]

@ywang96 actually if you check the nighly index: https://wheels.vllm.ai/nightly/vllm/

you found 12.9 only wheels available.