[Bug]: Load a custom model when VLLM_USE_V1=1

[akshay-loci]

Your current environment

INFO 01-28 23:42:17 __init__.py:183] Automatically detected platform cuda.
Collecting environment information...
PyTorch version: 2.5.1+cu124
Is debug build: False
CUDA used to build PyTorch: 12.4
ROCM used to build PyTorch: N/A

OS: Ubuntu 20.04.6 LTS (x86_64)
GCC version: (Ubuntu 9.4.0-1ubuntu1~20.04.2) 9.4.0
Clang version: Could not collect
CMake version: version 3.16.3
Libc version: glibc-2.31

Python version: 3.10.10 (main, Mar 21 2023, 18:45:11) [GCC 11.2.0] (64-bit runtime)
Python platform: Linux-5.15.0-1075-aws-x86_64-with-glibc2.31
Is CUDA available: True
CUDA runtime version: 12.1.105
CUDA_MODULE_LOADING set to: LAZY
GPU models and configuration: GPU 0: NVIDIA L4
Nvidia driver version: 535.230.02
cuDNN version: Probably one of the following:
/usr/lib/x86_64-linux-gnu/libcudnn.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_adv_infer.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_adv_train.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_infer.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_train.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_ops_infer.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_ops_train.so.8.9.0
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True

CPU:
Architecture:                         x86_64
CPU op-mode(s):                       32-bit, 64-bit
Byte Order:                           Little Endian
Address sizes:                        48 bits physical, 48 bits virtual
CPU(s):                               16
On-line CPU(s) list:                  0-15
Thread(s) per core:                   2
Core(s) per socket:                   8
Socket(s):                            1
NUMA node(s):                         1
Vendor ID:                            AuthenticAMD
CPU family:                           25
Model:                                1
Model name:                           AMD EPYC 7R13 Processor
Stepping:                             1
CPU MHz:                              2649.998
BogoMIPS:                             5299.99
Hypervisor vendor:                    KVM
Virtualization type:                  full
L1d cache:                            256 KiB
L1i cache:                            256 KiB
L2 cache:                             4 MiB
L3 cache:                             32 MiB
NUMA node0 CPU(s):                    0-15
Vulnerability Gather data sampling:   Not affected
Vulnerability Itlb multihit:          Not affected
Vulnerability L1tf:                   Not affected
Vulnerability Mds:                    Not affected
Vulnerability Meltdown:               Not affected
Vulnerability Mmio stale data:        Not affected
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed:               Not affected
Vulnerability Spec rstack overflow:   Mitigation; safe RET, no microcode
Vulnerability Spec store bypass:      Mitigation; Speculative Store Bypass disabled via prctl and seccomp
Vulnerability Spectre v1:             Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:             Mitigation; Retpolines; IBPB conditional; IBRS_FW; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected
Vulnerability Srbds:                  Not affected
Vulnerability Tsx async abort:        Not affected
Flags:                                fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc cpuid extd_apicid aperfmperf tsc_known_freq pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm cmp_legacy cr8_legacy abm sse4a misalignsse 3dnowprefetch topoext invpcid_single ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 invpcid rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 clzero xsaveerptr rdpru wbnoinvd arat npt nrip_save vaes vpclmulqdq rdpid

Versions of relevant libraries:
[pip3] mypy==1.13.0
[pip3] mypy-boto3-sagemaker-runtime==1.35.15
[pip3] mypy-extensions==1.0.0
[pip3] numpy==1.26.4
[pip3] nvidia-cublas-cu12==12.4.5.8
[pip3] nvidia-cuda-cupti-cu12==12.4.127
[pip3] nvidia-cuda-nvrtc-cu12==12.4.127
[pip3] nvidia-cuda-runtime-cu12==12.4.127
[pip3] nvidia-cudnn-cu12==9.1.0.70
[pip3] nvidia-cufft-cu12==11.2.1.3
[pip3] nvidia-curand-cu12==10.3.5.147
[pip3] nvidia-cusolver-cu12==11.6.1.9
[pip3] nvidia-cusparse-cu12==12.3.1.170
[pip3] nvidia-ml-py==12.560.30
[pip3] nvidia-nccl-cu12==2.21.5
[pip3] nvidia-nvjitlink-cu12==12.4.127
[pip3] nvidia-nvtx-cu12==12.4.127
[pip3] open-clip-torch==2.24.0
[pip3] pytorch-lightning==2.2.4
[pip3] pyzmq==26.0.3
[pip3] sentence-transformers==2.7.0
[pip3] torch==2.5.1
[pip3] torch.redstone==0.0.6
[pip3] torchmetrics==1.4.0.post0
[pip3] torchvision==0.20.1
[pip3] transformers==4.46.2
[pip3] triton==3.1.0
[conda] numpy                     1.26.4                   pypi_0    pypi
[conda] nvidia-cublas-cu12        12.4.5.8                 pypi_0    pypi
[conda] nvidia-cuda-cupti-cu12    12.4.127                 pypi_0    pypi
[conda] nvidia-cuda-nvrtc-cu12    12.4.127                 pypi_0    pypi
[conda] nvidia-cuda-runtime-cu12  12.4.127                 pypi_0    pypi
[conda] nvidia-cudnn-cu12         9.1.0.70                 pypi_0    pypi
[conda] nvidia-cufft-cu12         11.2.1.3                 pypi_0    pypi
[conda] nvidia-curand-cu12        10.3.5.147               pypi_0    pypi
[conda] nvidia-cusolver-cu12      11.6.1.9                 pypi_0    pypi
[conda] nvidia-cusparse-cu12      12.3.1.170               pypi_0    pypi
[conda] nvidia-ml-py              12.560.30                pypi_0    pypi
[conda] nvidia-nccl-cu12          2.21.5                   pypi_0    pypi
[conda] nvidia-nvjitlink-cu12     12.4.127                 pypi_0    pypi
[conda] nvidia-nvtx-cu12          12.4.127                 pypi_0    pypi
[conda] open-clip-torch           2.24.0                   pypi_0    pypi
[conda] pytorch-lightning         2.2.4                    pypi_0    pypi
[conda] pyzmq                     26.0.3                   pypi_0    pypi
[conda] sentence-transformers     2.7.0                    pypi_0    pypi
[conda] torch                     2.5.1                    pypi_0    pypi
[conda] torch-redstone            0.0.6                    pypi_0    pypi
[conda] torchmetrics              1.4.0.post0              pypi_0    pypi
[conda] torchvision               0.20.1                   pypi_0    pypi
[conda] transformers              4.46.2                   pypi_0    pypi
[conda] triton                    3.1.0                    pypi_0    pypi
ROCM Version: Could not collect
Neuron SDK Version: N/A
vLLM Version: 0.7.0
vLLM Build Flags:
CUDA Archs: 5.0;6.0;7.0;7.5;8.0;8.6;9.0; ROCm: Disabled; Neuron: Disabled
GPU Topology:
GPU0    CPU Affinity    NUMA Affinity   GPU NUMA ID
GPU0     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

NVIDIA_VISIBLE_DEVICES=all
NVIDIA_REQUIRE_CUDA=cuda>=12.1 brand=tesla,driver>=470,driver<471 brand=unknown,driver>=470,driver<471 brand=nvidia,driver>=470,driver<471 brand=nvidiartx,driver>=470,driver<471 brand=geforce,driver>=470,driver<471 brand=geforcertx,driver>=470,driver<471 brand=quadro,driver>=470,driver<471 brand=quadrortx,driver>=470,driver<471 brand=titan,driver>=470,driver<471 brand=titanrtx,driver>=470,driver<471 brand=tesla,driver>=525,driver<526 brand=unknown,driver>=525,driver<526 brand=nvidia,driver>=525,driver<526 brand=nvidiartx,driver>=525,driver<526 brand=geforce,driver>=525,driver<526 brand=geforcertx,driver>=525,driver<526 brand=quadro,driver>=525,driver<526 brand=quadrortx,driver>=525,driver<526 brand=titan,driver>=525,driver<526 brand=titanrtx,driver>=525,driver<526
TORCH_CUDA_ARCH_LIST=5.0;6.0;7.0;7.5;8.0;8.6;9.0
NCCL_VERSION=2.17.1-1
NVIDIA_DRIVER_CAPABILITIES=compute,graphics,utility,video
NVIDIA_PRODUCT_NAME=CUDA
NVIDIA_CUDA_END_OF_LIFE=1
CUDA_VERSION=12.1.0
CUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda
NVIDIA_DISABLE_REQUIRE=true
LD_LIBRARY_PATH=/system/conda/miniconda3/envs/cloudspace/lib/python3.10/site-packages/cv2/../../lib64:/opt/amazon/efa/lib:/opt/aws-ofi-nccl/lib:/usr/local/nvidia/lib:/usr/local/nvidia/lib64
MKL_THREADING_LAYER=GNU
NCCL_CUMEM_ENABLE=0
TORCHINDUCTOR_COMPILE_THREADS=1
CUDA_MODULE_LOADING=LAZY

Model Input Dumps

No response

🐛 Describe the bug

I have a custom VisionLanguage model that I'm running using VLLM.

Before loading it I register it like this:
ModelRegistry.register_model("VoxelPhiForCausalLM", VoxelForCausalLM)

and then load it like so:

llm = LLM(
    model="/teamspace/studios/this_studio/voxel_phi3_siglip",
    max_model_len=2048,
    disable_sliding_window=True,
    limit_mm_per_prompt={"image": 3},
    trust_remote_code=True,
    gpu_memory_utilization=0.8,
)

With VLLM 0.7.0 if I don't set VLLM_USE_V1=1 then it loads and runs fine but when setting it to 1 I get this error
Model architectures ["VoxelPhiForCausalLM"] are not supported for now.

Is it possible to use V1 architecture with custom models?

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.

[mgoin]

Can you share the full stacktrace? It is likely the real error is in the middle of the trace, and that message you shared is reporting the caught failure

[DarkLight1337]

VLMs don't support V1 by default. Please update your model by walking through the model development docs again. If that still doesn't work, can you share your model implementation?

[akshay-loci]

@mgoin @DarkLight1337 Looks like it was crashing because I was running it in a notebook, when running it from the command line it works fine. There's a very small difference in behavior though: When running generation without V1 enabled output.outputs[0].text does not contain the EOS token but with V1 enabled it does. I'm guessing this is not intentional?

[DarkLight1337]

This is similar to the problem I reported in #12504. cc @ywang96