Главная Обзор Помощь
Регистрация Вход
CRBC-MaaS-Platform-Project
/
LQDeployConfig
4
0
Ответвить 0
Файлы Задачи 0 Запросы на слияние 0 Вики
Ветка: master
Ветки Метки
LQDeployConfig
/
prod
/
models
/
minerU
/
tf_logs
/
minerU_transfer_2.logs

minerU_transfer_2.logs 12 KB
Постоянная ссылка История Исходник

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465 
  1. 

  2. ot@localhost minerU]# ls
    3. 

  3. compose.yaml  compose.yaml.bak  config  input  logs  minerU_transfer.logs  output  source
    4. 

  4. [root@localhost minerU]# time docker exec -it mineru-api mineru -p /app/input -o /app/output --source modelscope
    5. 

  5. 2026-03-21 22:09:23.104 | INFO     | mineru.utils.engine_utils:get_vlm_engine:32 - Using vllm-engine as the inference engine for VLM.
    6. 

  6. Downloading Model from https://www.modelscope.cn to directory: /root/.cache/modelscope/models/OpenDataLab/MinerU2.5-2509-1.2B
    7. 

  7. 2026-03-21 22:09:28,881 - modelscope - INFO - Target directory already exists, skipping creation.
    8. 

  8. 2026-03-21 22:09:28.881 | INFO     | mineru.backend.vlm.utils:enable_custom_logits_processors:55 - compute_capability: 9.0 >= 8.0 and vllm version: 0.10.1.1 >= 0.10.1, enable custom_logits_processors
    9. 

  9. /usr/local/lib/python3.12/dist-packages/transformers/utils/hub.py:111: FutureWarning: Using `TRANSFORMERS_CACHE` is deprecated and will be removed in v5 of Transformers. Use `HF_HOME` instead.
    10. 

  10.   warnings.warn(
    11. 

  11. INFO 03-21 22:09:29 [__init__.py:241] Automatically detected platform cuda.
    12. 

  12. INFO 03-21 22:09:30 [utils.py:326] non-default args: {'model': '/root/.cache/modelscope/models/OpenDataLab/MinerU2___5-2509-1___2B', 'gpu_memory_utilization': 0.5, 'disable_log_stats': True, 'logits_processors': [<class 'mineru_vl_utils.logits_processor.vllm_v1_no_repeat_ngram.VllmV1NoRepeatNGramLogitsProcessor'>]}
    13. 

  13. INFO 03-21 22:09:36 [__init__.py:711] Resolved architecture: Qwen2VLForConditionalGeneration
    14. 

  14. INFO 03-21 22:09:36 [__init__.py:1750] Using max model len 16384
    15. 

  15. INFO 03-21 22:09:37 [scheduler.py:222] Chunked prefill is enabled with max_num_batched_tokens=16384.
    16. 

  16. WARNING 03-21 22:09:38 [__init__.py:2921] We must use the `spawn` multiprocessing start method. Overriding VLLM_WORKER_MULTIPROC_METHOD to 'spawn'. See https://docs.vllm.ai/en/latest/usage/troubleshooting.html#python-multiprocessing for more information. Reasons: CUDA is initialized
    17. 

  17. /usr/local/lib/python3.12/dist-packages/transformers/utils/hub.py:111: FutureWarning: Using `TRANSFORMERS_CACHE` is deprecated and will be removed in v5 of Transformers. Use `HF_HOME` instead.
    18. 

  18.   warnings.warn(
    19. 

  19. INFO 03-21 22:09:42 [__init__.py:241] Automatically detected platform cuda.
    20. 

  20. (EngineCore_0 pid=1106) INFO 03-21 22:09:43 [core.py:636] Waiting for init message from front-end.
    21. 

  21. (EngineCore_0 pid=1106) INFO 03-21 22:09:43 [core.py:74] Initializing a V1 LLM engine (v0.10.1.1) with config: model='/root/.cache/modelscope/models/OpenDataLab/MinerU2___5-2509-1___2B', speculative_config=None, tokenizer='/root/.cache/modelscope/models/OpenDataLab/MinerU2___5-2509-1___2B', skip_tokenizer_init=False, tokenizer_mode=auto, revision=None, override_neuron_config={}, tokenizer_revision=None, trust_remote_code=False, dtype=torch.bfloat16, max_seq_len=16384, download_dir=None, load_format=auto, tensor_parallel_size=1, pipeline_parallel_size=1, disable_custom_all_reduce=False, quantization=None, enforce_eager=False, kv_cache_dtype=auto, device_config=cuda, decoding_config=DecodingConfig(backend='auto', disable_fallback=False, disable_any_whitespace=False, disable_additional_properties=False, reasoning_backend=''), observability_config=ObservabilityConfig(show_hidden_metrics_for_version=None, otlp_traces_endpoint=None, collect_detailed_traces=None), seed=0, served_model_name=/root/.cache/modelscope/models/OpenDataLab/MinerU2___5-2509-1___2B, enable_prefix_caching=True, chunked_prefill_enabled=True, use_async_output_proc=True, pooler_config=None, compilation_config={"level":3,"debug_dump_path":"","cache_dir":"","backend":"","custom_ops":[],"splitting_ops":["vllm.unified_attention","vllm.unified_attention_with_output","vllm.mamba_mixer2"],"use_inductor":true,"compile_sizes":[],"inductor_compile_config":{"enable_auto_functionalized_v2":false},"inductor_passes":{},"cudagraph_mode":1,"use_cudagraph":true,"cudagraph_num_of_warmups":1,"cudagraph_capture_sizes":[512,504,496,488,480,472,464,456,448,440,432,424,416,408,400,392,384,376,368,360,352,344,336,328,320,312,304,296,288,280,272,264,256,248,240,232,224,216,208,200,192,184,176,168,160,152,144,136,128,120,112,104,96,88,80,72,64,56,48,40,32,24,16,8,4,2,1],"cudagraph_copy_inputs":false,"full_cuda_graph":false,"pass_config":{},"max_capture_size":512,"local_cache_dir":null}
    22. 

  22. (EngineCore_0 pid=1106) INFO 03-21 22:09:45 [parallel_state.py:1134] rank 0 in world size 1 is assigned as DP rank 0, PP rank 0, TP rank 0, EP rank 0
    23. 

  23. (EngineCore_0 pid=1106) INFO 03-21 22:09:46 [topk_topp_sampler.py:50] Using FlashInfer for top-p & top-k sampling.
    24. 

  24. (EngineCore_0 pid=1106) INFO 03-21 22:09:47 [gpu_model_runner.py:1953] Starting to load model /root/.cache/modelscope/models/OpenDataLab/MinerU2___5-2509-1___2B...
    25. 

  25. (EngineCore_0 pid=1106) INFO 03-21 22:09:48 [gpu_model_runner.py:1985] Loading model from scratch...
    26. 

  26. (EngineCore_0 pid=1106) WARNING 03-21 22:09:48 [cuda.py:211] Current `vllm-flash-attn` has a bug inside vision module, so we use xformers backend instead. You can run `pip install flash-attn` to use flash-attention backend.
    27. 

  27. (EngineCore_0 pid=1106) INFO 03-21 22:09:48 [cuda.py:328] Using Flash Attention backend on V1 engine.
    28. 

  28. Loading safetensors checkpoint shards:   0% Completed | 0/1 [00:00<?, ?it/s]
    29. 

  29. Loading safetensors checkpoint shards: 100% Completed | 1/1 [00:00<00:00,  1.29it/s]
    30. 

  30. Loading safetensors checkpoint shards: 100% Completed | 1/1 [00:00<00:00,  1.29it/s]
    31. 

  31. (EngineCore_0 pid=1106) 
    32. 

  32. (EngineCore_0 pid=1106) INFO 03-21 22:09:49 [default_loader.py:262] Loading weights took 0.81 seconds
    33. 

  33. (EngineCore_0 pid=1106) INFO 03-21 22:09:49 [gpu_model_runner.py:2007] Model loading took 2.1647 GiB and 0.951355 seconds
    34. 

  34. (EngineCore_0 pid=1106) INFO 03-21 22:09:49 [gpu_model_runner.py:2591] Encoder cache will be initialized with a budget of 16384 tokens, and profiled with 1 video items of the maximum feature size.
    35. 

  35. (EngineCore_0 pid=1106) INFO 03-21 22:09:56 [backends.py:548] Using cache directory: /root/.cache/vllm/torch_compile_cache/b19ac70222/rank_0_0/backbone for vLLM's torch.compile
    36. 

  36. (EngineCore_0 pid=1106) INFO 03-21 22:09:56 [backends.py:559] Dynamo bytecode transform time: 2.95 s
    37. 

  37. (EngineCore_0 pid=1106) INFO 03-21 22:09:59 [backends.py:161] Directly load the compiled graph(s) for dynamic shape from the cache, took 2.552 s
    38. 

  38. (EngineCore_0 pid=1106) INFO 03-21 22:09:59 [monitor.py:34] torch.compile takes 2.95 s in total
    39. 

  39. (EngineCore_0 pid=1106) /usr/local/lib/python3.12/dist-packages/torch/utils/cpp_extension.py:2356: UserWarning: TORCH_CUDA_ARCH_LIST is not set, all archs for visible cards are included for compilation. 
    40. 

  40. (EngineCore_0 pid=1106) If this is not desired, please set os.environ['TORCH_CUDA_ARCH_LIST'].
    41. 

  41. (EngineCore_0 pid=1106)   warnings.warn(
    42. 

  42. (EngineCore_0 pid=1106) INFO 03-21 22:10:00 [gpu_worker.py:276] Available KV cache memory: 65.01 GiB
    43. 

  43. (EngineCore_0 pid=1106) INFO 03-21 22:10:00 [kv_cache_utils.py:849] GPU KV cache size: 5,680,496 tokens
    44. 

  44. (EngineCore_0 pid=1106) INFO 03-21 22:10:00 [kv_cache_utils.py:853] Maximum concurrency for 16,384 tokens per request: 346.71x
    45. 

  45. Capturing CUDA graphs (mixed prefill-decode, PIECEWISE): 100%|███████████████████████████████████████████████████████████████| 67/67 [00:02<00:00, 30.70it/s](EngineCore_0 pid=1106) INFO 03-21 22:10:02 [gpu_model_runner.py:2708] Graph capturing finished in 2 secs, took 0.54 GiB
    46. 

  46. (EngineCore_0 pid=1106) INFO 03-21 22:10:02 [core.py:214] init engine (profile, create kv cache, warmup model) took 13.45 seconds
    47. 

  47. INFO 03-21 22:10:04 [llm.py:298] Supported_tasks: ['generate']
    48. 

  48. 2026-03-21 22:10:04.249 | INFO     | mineru.backend.vlm.vlm_analyze:get_model:218 - get vllm-engine predictor cost: 36.48s
    49. 

  49. 2026-03-21 22:10:05.955 | INFO     | mineru.backend.hybrid.hybrid_analyze:get_batch_ratio:365 - hybrid batch ratio (auto, vram=140GB): 16
    50. 

  50. Adding requests: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████| 33/33 [00:02<00:00, 13.69it/s]Processed prompts: 100%|██████████████████████████████████████████| 33/33 [00:04<00:00,  7.28it/s, est. speed input: 10152.20 toks/s, output: 1174.95 toks/s]Downloading Model from https://www.modelscope.cn to directory: /root/.cache/modelscope/models/OpenDataLab/PDF-Extract-Kit-1.0
    51. 

  51. 2026-03-21 22:10:15,718 - modelscope - INFO - Target directory already exists, skipping creation.
    52. 

  52. Downloading Model from https://www.modelscope.cn to directory: /root/.cache/modelscope/models/OpenDataLab/PDF-Extract-Kit-1.0
    53. 

  53. 2026-03-21 22:10:16,976 - modelscope - INFO - Target directory already exists, skipping creation.
    54. 

  54. Downloading Model from https://www.modelscope.cn to directory: /root/.cache/modelscope/models/OpenDataLab/PDF-Extract-Kit-1.0
    55. 

  55. 2026-03-21 22:10:19,464 - modelscope - INFO - Target directory already exists, skipping creation.
    56. 

  56. Downloading Model from https://www.modelscope.cn to directory: /root/.cache/modelscope/models/OpenDataLab/PDF-Extract-Kit-1.0
    57. 

  57. 2026-03-21 22:10:20,849 - modelscope - INFO - Target directory already exists, skipping creation.
    58. 

  58. MFD Predict: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████| 33/33 [00:01<00:00, 20.28it/s]MFR Predict: 0it [00:00, ?it/s]
    59. 

  59. OCR-det: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████| 24/24 [00:03<00:00,  7.24it/s]2026-03-21 22:10:28.913 | INFO     | mineru.cli.common:_process_output:168 - local output dir is /app/output/83《中华人民共和国突发事件应对法》（主席令第69号（2019修订））/hybrid_auto
    60. 

  60. Adding requests: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████| 4/4 [00:00<00:00, 36.38it/s]Processed prompts: 100%|██████████████████████████████████████████████| 4/4 [00:02<00:00,  1.64it/s, est. speed input: 2282.62 toks/s, output: 713.52 toks/s]Adding requests: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████| 76/76 [00:00<00:00, 221.55it/s]Processed prompts: 100%|███████████████████████████████████████████| 76/76 [00:01<00:00, 69.82it/s, est. speed input: 4371.88 toks/s, output: 1661.26 toks/s]2026-03-21 22:10:33.850 | INFO     | mineru.cli.common:_process_output:168 - local output dir is /app/output/33《铸钢砂》（YBT5150-93）/hybrid_auto
    61. 

  61. ERROR 03-21 22:10:35 [core_client.py:562] Engine core proc EngineCore_0 died unexpectedly, shutting down client.
    62. 

  62. 

  63. real    1m17.713s
    64. 

  64. user    0m0.048s
    65. 

  65. sys     0m0.044s
    66.