صفحهٔ اصلی گشت‌و‌گذار راهنما
ثبت نام ورود
CRBC-MaaS-Platform-Project
/
LQDeployConfig
4
0
انشعاب 0
پرونده‌ها مشکلات 0 درخواست واکشی 0 ویکی
درخت: 6800f79aef
شاخه‌ها تگ‌ها
LQDeployConfig
/
prod
/
models
/
minerU
/
tf_logs
/
minerU_transfer_3.logs

minerU_transfer_3.logs 12 KB
تاريخچه خام

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667 
  1. 

  2. ot@localhost input]# cd ..
    3. 

  3. [root@localhost minerU]# cd input/
    4. 

  4. [root@localhost input]# ls
    5. 

  5. fa56f7a8-bd36-4140-8cb1-f9f973ce8745.pdf
    6. 

  6. [root@localhost input]# du -sh *
    7. 

  7. 22M     fa56f7a8-bd36-4140-8cb1-f9f973ce8745.pdf
    8. 

  8. [root@localhost input]# cd ..
    9. 

  9. [root@localhost minerU]# time docker exec -it mineru-api mineru -p /app/input -o /app/output --source modelscope
    10. 

  10. 2026-03-21 22:16:57.538 | INFO     | mineru.utils.engine_utils:get_vlm_engine:32 - Using vllm-engine as the inference engine for VLM.
    11. 

  11. Downloading Model from https://www.modelscope.cn to directory: /root/.cache/modelscope/models/OpenDataLab/MinerU2.5-2509-1.2B
    12. 

  12. 2026-03-21 22:17:00,921 - modelscope - INFO - Target directory already exists, skipping creation.
    13. 

  13. 2026-03-21 22:17:00.922 | 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
    14. 

  14. /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.
    15. 

  15.   warnings.warn(
    16. 

  16. INFO 03-21 22:17:02 [__init__.py:241] Automatically detected platform cuda.
    17. 

  17. INFO 03-21 22:17:03 [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'>]}
    18. 

  18. INFO 03-21 22:17:09 [__init__.py:711] Resolved architecture: Qwen2VLForConditionalGeneration
    19. 

  19. INFO 03-21 22:17:09 [__init__.py:1750] Using max model len 16384
    20. 

  20. INFO 03-21 22:17:09 [scheduler.py:222] Chunked prefill is enabled with max_num_batched_tokens=16384.
    21. 

  21. WARNING 03-21 22:17:11 [__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
    22. 

  22. /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.
    23. 

  23.   warnings.warn(
    24. 

  24. INFO 03-21 22:17:16 [__init__.py:241] Automatically detected platform cuda.
    25. 

  25. (EngineCore_0 pid=1533) INFO 03-21 22:17:17 [core.py:636] Waiting for init message from front-end.
    26. 

  26. (EngineCore_0 pid=1533) INFO 03-21 22:17:17 [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}
    27. 

  27. (EngineCore_0 pid=1533) INFO 03-21 22:17:19 [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
    28. 

  28. (EngineCore_0 pid=1533) INFO 03-21 22:17:19 [topk_topp_sampler.py:50] Using FlashInfer for top-p & top-k sampling.
    29. 

  29. (EngineCore_0 pid=1533) INFO 03-21 22:17:21 [gpu_model_runner.py:1953] Starting to load model /root/.cache/modelscope/models/OpenDataLab/MinerU2___5-2509-1___2B...
    30. 

  30. (EngineCore_0 pid=1533) INFO 03-21 22:17:22 [gpu_model_runner.py:1985] Loading model from scratch...
    31. 

  31. (EngineCore_0 pid=1533) WARNING 03-21 22:17:22 [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.
    32. 

  32. (EngineCore_0 pid=1533) INFO 03-21 22:17:22 [cuda.py:328] Using Flash Attention backend on V1 engine.
    33. 

  33. Loading safetensors checkpoint shards:   0% Completed | 0/1 [00:00<?, ?it/s]
    34. 

  34. Loading safetensors checkpoint shards: 100% Completed | 1/1 [00:00<00:00,  1.11it/s]
    35. 

  35. Loading safetensors checkpoint shards: 100% Completed | 1/1 [00:00<00:00,  1.11it/s]
    36. 

  36. (EngineCore_0 pid=1533) 
    37. 

  37. (EngineCore_0 pid=1533) INFO 03-21 22:17:23 [default_loader.py:262] Loading weights took 0.94 seconds
    38. 

  38. (EngineCore_0 pid=1533) INFO 03-21 22:17:23 [gpu_model_runner.py:2007] Model loading took 2.1647 GiB and 1.094801 seconds
    39. 

  39. (EngineCore_0 pid=1533) INFO 03-21 22:17:23 [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.
    40. 

  40. (EngineCore_0 pid=1533) INFO 03-21 22:17:30 [backends.py:548] Using cache directory: /root/.cache/vllm/torch_compile_cache/b19ac70222/rank_0_0/backbone for vLLM's torch.compile
    41. 

  41. (EngineCore_0 pid=1533) INFO 03-21 22:17:30 [backends.py:559] Dynamo bytecode transform time: 2.96 s
    42. 

  42. (EngineCore_0 pid=1533) INFO 03-21 22:17:33 [backends.py:161] Directly load the compiled graph(s) for dynamic shape from the cache, took 2.876 s
    43. 

  43. (EngineCore_0 pid=1533) INFO 03-21 22:17:34 [monitor.py:34] torch.compile takes 2.96 s in total
    44. 

  44. (EngineCore_0 pid=1533) /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. 
    45. 

  45. (EngineCore_0 pid=1533) If this is not desired, please set os.environ['TORCH_CUDA_ARCH_LIST'].
    46. 

  46. (EngineCore_0 pid=1533)   warnings.warn(
    47. 

  47. (EngineCore_0 pid=1533) INFO 03-21 22:17:34 [gpu_worker.py:276] Available KV cache memory: 65.01 GiB
    48. 

  48. (EngineCore_0 pid=1533) INFO 03-21 22:17:35 [kv_cache_utils.py:849] GPU KV cache size: 5,680,496 tokens
    49. 

  49. (EngineCore_0 pid=1533) INFO 03-21 22:17:35 [kv_cache_utils.py:853] Maximum concurrency for 16,384 tokens per request: 346.71x
    50. 

  50. Capturing CUDA graphs (mixed prefill-decode, PIECEWISE): 100%|███████████████████████████████████████████████████████████████| 67/67 [00:04<00:00, 14.88it/s](EngineCore_0 pid=1533) INFO 03-21 22:17:39 [gpu_model_runner.py:2708] Graph capturing finished in 5 secs, took 0.54 GiB
    51. 

  51. (EngineCore_0 pid=1533) INFO 03-21 22:17:39 [core.py:214] init engine (profile, create kv cache, warmup model) took 16.14 seconds
    52. 

  52. INFO 03-21 22:17:41 [llm.py:298] Supported_tasks: ['generate']
    53. 

  53. 2026-03-21 22:17:41.167 | INFO     | mineru.backend.vlm.vlm_analyze:get_model:218 - get vllm-engine predictor cost: 41.47s
    54. 

  54. 2026-03-21 22:17:51.118 | INFO     | mineru.backend.hybrid.hybrid_analyze:get_batch_ratio:365 - hybrid batch ratio (auto, vram=140GB): 16
    55. 

  55. Adding requests: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████| 188/188 [00:04<00:00, 38.89it/s]Processed prompts: 100%|█████████████████████████████████████████| 188/188 [00:30<00:00,  6.26it/s, est. speed input: 8730.38 toks/s, output: 3129.15 toks/s]Adding requests: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████| 154/154 [00:01<00:00, 103.19it/s]Processed prompts: 100%|█████████████████████████████████████████| 154/154 [00:13<00:00, 11.23it/s, est. speed input: 5665.97 toks/s, output: 2169.77 toks/s]Downloading Model from https://www.modelscope.cn to directory: /root/.cache/modelscope/models/OpenDataLab/PDF-Extract-Kit-1.0
    56. 

  56. 2026-03-21 22:18:52,003 - modelscope - INFO - Target directory already exists, skipping creation.
    57. 

  57. Downloading Model from https://www.modelscope.cn to directory: /root/.cache/modelscope/models/OpenDataLab/PDF-Extract-Kit-1.0
    58. 

  58. 2026-03-21 22:18:53,218 - modelscope - INFO - Target directory already exists, skipping creation.
    59. 

  59. Downloading Model from https://www.modelscope.cn to directory: /root/.cache/modelscope/models/OpenDataLab/PDF-Extract-Kit-1.0
    60. 

  60. 2026-03-21 22:18:55,660 - modelscope - INFO - Target directory already exists, skipping creation.
    61. 

  61. Downloading Model from https://www.modelscope.cn to directory: /root/.cache/modelscope/models/OpenDataLab/PDF-Extract-Kit-1.0
    62. 

  62. 2026-03-21 22:18:57,083 - modelscope - INFO - Target directory already exists, skipping creation.
    63. 

  63. MFD Predict: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████| 188/188 [00:07<00:00, 24.98it/s]MFR Predict: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████| 635/635 [00:03<00:00, 162.59it/s]OCR-det: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████| 56/56 [00:14<00:00,  3.86it/s]OCR-rec Predict: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████| 2226/2226 [00:04<00:00, 502.40it/s]2026-03-21 22:19:46.740 | INFO     | mineru.cli.common:_process_output:168 - local output dir is /app/output/fa56f7a8-bd36-4140-8cb1-f9f973ce8745/hybrid_autoERROR 03-21 22:19:48 [core_client.py:562] Engine core proc EngineCore_0 died unexpectedly, shutting down client.
    64. 

  64. 

  65. real    2m56.204s
    66. 

  66. user    0m0.054s
    67. 

  67. Sys     0m0.057s
    68.