Multiple errors while following steps in docker_pytorch_inference_gpu.md

[js333031]

Please see attached of console logs while following steps described in this.

Host system is Ubuntu 24.04.2 LTS, 8th Gen Core i7 with 48G DDR4, A770 LE GPU

Line 207: ModuleNotFoundError: No module named 'omegaconf'
Lines 304 onwards

ipex-llm-container-errors.txt

[cyita]

Hi @js333031, you need to pip install omegaconf pandas and try to downgrade transformers pip install transformers==4.44.2.

[js333031]

After installing the suggested transformers package in previous comment, following error is encountered: ModuleNotFoundError: No module named 'transformers.models.mllama'

Full trace below. Thanks.

root@clftower:/benchmark/all-in-one# source ipex-llm-init --gpu --device Arc
found intel-openmp in /usr/local/lib/libiomp5.so
found oneapi in /opt/intel/oneapi/setvars.sh

:: initializing oneAPI environment ...
   bash: BASH_VERSION = 5.1.16(1)-release
   args: Using "$@" for setvars.sh arguments: --force
:: advisor -- latest
:: ccl -- latest
:: compiler -- latest
:: dal -- latest
:: debugger -- latest
:: dev-utilities -- latest
:: dnnl -- latest
:: dpcpp-ct -- latest
:: dpl -- latest
:: ipp -- latest
:: ippcp -- latest
:: mkl -- latest
:: mpi -- latest
:: pti -- latest
:: tbb -- latest
:: umf -- latest
:: vtune -- latest
:: oneAPI environment initialized ::

[W626 21:40:15.068372934 OperatorEntry.cpp:154] Warning: Warning only once for all operators,  other operators may also be overridden.
  Overriding a previously registered kernel for the same operator and the same dispatch key
  operator: aten::_validate_compressed_sparse_indices(bool is_crow, Tensor compressed_idx, Tensor plain_idx, int cdim, int dim, int nnz) -> ()
    registered at /pytorch/build/aten/src/ATen/RegisterSchema.cpp:6
  dispatch key: XPU
  previous kernel: registered at /pytorch/build/aten/src/ATen/RegisterCPU.cpp:30477
       new kernel: registered at /build/intel-pytorch-extension/build/Release/csrc/gpu/csrc/aten/generated/ATen/RegisterXPU.cpp:468 (function operator())
[W626 21:40:17.419380914 OperatorEntry.cpp:154] Warning: Warning only once for all operators,  other operators may also be overridden.
  Overriding a previously registered kernel for the same operator and the same dispatch key
  operator: aten::_validate_compressed_sparse_indices(bool is_crow, Tensor compressed_idx, Tensor plain_idx, int cdim, int dim, int nnz) -> ()
    registered at /pytorch/build/aten/src/ATen/RegisterSchema.cpp:6
  dispatch key: XPU
  previous kernel: registered at /pytorch/build/aten/src/ATen/RegisterCPU.cpp:30477
       new kernel: registered at /build/intel-pytorch-extension/build/Release/csrc/gpu/csrc/aten/generated/ATen/RegisterXPU.cpp:468 (function operator())
+++++ Env Variables +++++
Internal:
    ENABLE_IOMP     = 1
    ENABLE_GPU      = 1
    ENABLE_JEMALLOC = 0
    ENABLE_TCMALLOC = 0
    LIB_DIR    = /usr/local/lib
    BIN_DIR    = bin64
    LLM_DIR    = /usr/local/lib/python3.11/dist-packages/ipex_llm

Exported:
    LD_PRELOAD             = /usr/local/lib/libiomp5.so
    OMP_NUM_THREADS        = 6
    MALLOC_CONF            =
    USE_XETLA              = OFF
    ENABLE_SDP_FUSION      =
    SYCL_CACHE_PERSISTENT  = 1
    BIGDL_LLM_XMX_DISABLED =
    SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS = 1
+++++++++++++++++++++++++
Complete.
root@clftower:/benchmark/all-in-one#

root@clftower:/benchmark/all-in-one# python run.py
[W626 21:41:13.933371495 OperatorEntry.cpp:154] Warning: Warning only once for all operators,  other operators may also be overridden.
  Overriding a previously registered kernel for the same operator and the same dispatch key
  operator: aten::_validate_compressed_sparse_indices(bool is_crow, Tensor compressed_idx, Tensor plain_idx, int cdim, int dim, int nnz) -> ()
    registered at /pytorch/build/aten/src/ATen/RegisterSchema.cpp:6
  dispatch key: XPU
  previous kernel: registered at /pytorch/build/aten/src/ATen/RegisterCPU.cpp:30477
       new kernel: registered at /build/intel-pytorch-extension/build/Release/csrc/gpu/csrc/aten/generated/ATen/RegisterXPU.cpp:468 (function operator())
Loading checkpoint shards:   0%|                                                                                                                                                              | 0/4 [00:00<?, ?it/s][W626 21:41:16.593663798 OperatorEntry.cpp:154] Warning: Warning only once for all operators,  other operators may also be overridden.
  Overriding a previously registered kernel for the same operator and the same dispatch key
  operator: aten::_validate_compressed_sparse_indices(bool is_crow, Tensor compressed_idx, Tensor plain_idx, int cdim, int dim, int nnz) -> ()
    registered at /pytorch/build/aten/src/ATen/RegisterSchema.cpp:6
  dispatch key: XPU
  previous kernel: registered at /pytorch/build/aten/src/ATen/RegisterCPU.cpp:30477
       new kernel: registered at /build/intel-pytorch-extension/build/Release/csrc/gpu/csrc/aten/generated/ATen/RegisterXPU.cpp:468 (function operator())
Loading checkpoint shards: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 4/4 [00:15<00:00,  3.86s/it]
2025-06-26 21:41:31,468 - ipex_llm.transformers.utils - WARNING - sym_int4 is deprecated, use woq_int4 instead, if you are loading saved sym_int4 low bit model, please resaved it with woq_int4
2025-06-26 21:41:31,469 - ipex_llm.transformers.utils - INFO - Converting the current model to sym_int4 format......
/usr/local/lib/python3.11/dist-packages/torch/nn/init.py:511: UserWarning: Initializing zero-element tensors is a no-op
  warnings.warn("Initializing zero-element tensors is a no-op")
INFO 06-26 21:41:33 [__init__.py:239] Automatically detected platform xpu.
Traceback (most recent call last):
  File "/benchmark/all-in-one/run.py", line 2338, in <module>
    run_model(model, api, in_out_pairs, conf['local_model_hub'], conf['warm_up'], conf['num_trials'], conf['num_beams'],
  File "/benchmark/all-in-one/run.py", line 157, in run_model
    result = run_transformer_int4_gpu(repo_id, local_model_hub, in_out_pairs, warm_up, num_trials, num_beams, low_bit, batch_size, cpu_embedding, fp16=True, lookahead=lookahead, task=task)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/benchmark/all-in-one/run.py", line 538, in run_transformer_int4_gpu
    model = AutoModelForCausalLM.from_pretrained(model_path, optimize_model=True, load_in_low_bit=low_bit,
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.11/unittest/mock.py", line 1378, in patched
    return func(*newargs, **newkeywargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/ipex_llm/transformers/model.py", line 349, in from_pretrained
    model = cls.load_convert(q_k, optimize_model, *args, **kwargs)
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/ipex_llm/transformers/model.py", line 502, in load_convert
    model = ggml_convert_low_bit(model, qtype, optimize_model,
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/ipex_llm/transformers/convert.py", line 1145, in ggml_convert_low_bit
    model, has_been_replaced = _replace_with_low_bit_linear(
                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/ipex_llm/transformers/convert.py", line 489, in _replace_with_low_bit_linear
    is_linear, linear_args = is_linear_module(module)
                             ^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/dist-packages/ipex_llm/transformers/convert.py", line 152, in is_linear_module
    from vllm.model_executor.layers.linear import (
  File "/usr/local/lib/python3.11/dist-packages/vllm-0.8.3+ipexllm.xpu-py3.11-linux-x86_64.egg/vllm/__init__.py", line 12, in <module>
    from vllm.engine.arg_utils import AsyncEngineArgs, EngineArgs
  File "/usr/local/lib/python3.11/dist-packages/vllm-0.8.3+ipexllm.xpu-py3.11-linux-x86_64.egg/vllm/engine/arg_utils.py", line 15, in <module>
    from vllm.config import (CacheConfig, CompilationConfig, ConfigFormat,
  File "/usr/local/lib/python3.11/dist-packages/vllm-0.8.3+ipexllm.xpu-py3.11-linux-x86_64.egg/vllm/config.py", line 35, in <module>
    from vllm.transformers_utils.config import (
  File "/usr/local/lib/python3.11/dist-packages/vllm-0.8.3+ipexllm.xpu-py3.11-linux-x86_64.egg/vllm/transformers_utils/config.py", line 31, in <module>
    from vllm.transformers_utils.configs import (ChatGLMConfig, Cohere2Config,
  File "/usr/local/lib/python3.11/dist-packages/vllm-0.8.3+ipexllm.xpu-py3.11-linux-x86_64.egg/vllm/transformers_utils/configs/__init__.py", line 18, in <module>
    from vllm.transformers_utils.configs.mllama import MllamaConfig
  File "/usr/local/lib/python3.11/dist-packages/vllm-0.8.3+ipexllm.xpu-py3.11-linux-x86_64.egg/vllm/transformers_utils/configs/mllama.py", line 3, in <module>
    from transformers.models.mllama import configuration_mllama as mllama_hf_config
ModuleNotFoundError: No module named 'transformers.models.mllama'
root@clftower:/benchmark/all-in-one# 

[qiyuangong]

This error is caused by an incompatible transformer version. Seems you are trying to run https://huggingface.co/meta-llama/Llama-3.2-11B-Vision-Instruct

Can you share your model type or config.yaml?

[js333031]

I'm using Llama-3.1-8B-Instruct. Here's my config.yaml:

root@clftower:/benchmark/all-in-one# cat config.yaml
repo_id:
  # - 'THUDM/chatglm2-6b'
  # - 'meta-llama/Llama-2-7b-chat-hf'
  - 'meta-llama/Llama-3.1-8B-Instruct'
  # - 'liuhaotian/llava-v1.5-7b' # requires a LLAVA_REPO_DIR env variables pointing to the llava dir; added only for gpu win related test_api now
local_model_hub: '/llm/models'
warm_up: 1 # must set >=2 when run "pipeline_parallel_gpu" test_api
num_trials: 3
num_beams: 1 # default to greedy search
low_bit: 'sym_int4' # default to use 'sym_int4' (i.e. symmetric int4)
batch_size: 1 # default to 1
in_out_pairs:
  - '32-32'
  - '960-64'
  - '1024-128'
test_api:
  - "transformer_int4_fp16_gpu"             # on Intel GPU, transformer-like API, (qtype=int4), (dtype=fp16)
  # - "transformer_int4_fp16_gpu_win"       # on Intel GPU for Windows, transformer-like API, (qtype=int4), (dtype=fp16)
  # - "transformer_int4_gpu"                # on Intel GPU, transformer-like API, (qtype=int4), (dtype=fp32)
  # - "transformer_int4_gpu_win"            # on Intel GPU for Windows, transformer-like API, (qtype=int4), (dtype=fp32)
  # - "transformer_int4_loadlowbit_gpu_win" # on Intel GPU for Windows, transformer-like API, (qtype=int4), use load_low_bit API. Please make sure you have used the save.py to save the converted low bit model
  # - "transformer_int4_fp16_loadlowbit_gpu_win" # on Intel GPU for Windows, transformer-like API, (qtype=int4), (dtype=fp16), use load_low_bit API. Please make sure you have used the save.py to save the converted low bit model
  # - "bigdl_fp16_gpu"                      # on Intel GPU, use ipex-llm transformers API, (dtype=fp16), (qtype=fp16)
  # - "optimize_model_gpu"                  # on Intel GPU, can optimize any pytorch models include transformer model
  # - "deepspeed_optimize_model_gpu"        # on Intel GPU, deepspeed autotp inference
  # - "pipeline_parallel_gpu"               # on Intel GPU, pipeline parallel inference
  # - "speculative_gpu"                     # on Intel GPU, inference with self-speculative decoding
  # - "transformer_int4"                    # on Intel CPU, transformer-like API, (qtype=int4)
  # - "native_int4"                         # on Intel CPU
  # - "optimize_model"                      # on Intel CPU, can optimize any pytorch models include transformer model
  # - "pytorch_autocast_bf16"               # on Intel CPU
  # - "transformer_autocast_bf16"           # on Intel CPU
  # - "bigdl_ipex_bf16"                     # on Intel CPU, (qtype=bf16)
  # - "bigdl_ipex_int4"                     # on Intel CPU, (qtype=int4)
  # - "bigdl_ipex_int8"                     # on Intel CPU, (qtype=int8)
  # - "speculative_cpu"                     # on Intel CPU, inference with self-speculative decoding
  # - "deepspeed_transformer_int4_cpu"      # on Intel CPU, deepspeed autotp inference
  # - "transformers_int4_npu_win"           # on Intel NPU for Windows,  transformer-like API, (qtype=int4)
  # - "transformers_int4_loadlowbit_npu_win" # on Intel NPU for Windows, transformer-like API, (qtype=int4), use load_low_bit API. Please make sure you have used the save_npu.py to save the converted low bit model
  # - "transformers_int4_npu_pipeline_win"  # on Intel NPU for Windows,  transformer-like API, (qtype=int4)
cpu_embedding: False # whether put embedding to CPU
streaming: False # whether output in streaming way (only available now for gpu win related test_api)
optimize_model: True # whether apply further optimization on NPU (only available now for transformers_int4_npu_win test_api)
use_fp16_torch_dtype: True # whether use fp16 for non-linear layer (only available now for "pipeline_parallel_gpu" test_api)
task: 'continuation' # task can be 'continuation', 'QA' and 'summarize'
transpose_value_cache: True # whether apply transposed v_cache optimization on NPU (only available now for transformers_int4_npu_win test_api)
npu_group_size: 0 # this can only be either 0 or 128, and only works for `transformers_int4_npu_win` / `transformers_int4_npu_pipeline_win`
root@clftower:/benchmark/all-in-one#