PP-StrucutreV3初始化时报错

[aa851182472]

---

C++ Traceback (most recent call last):

No stack trace in paddle, may be caused by external reasons.

---

Error Message Summary:

FatalError: Segmentation fault is detected by the operating system.
[TimeInfo: *** Aborted at 1759027014 (unix time) try "date -d @1759027014" if you are using GNU date ***]
[SignalInfo: *** SIGSEGV (@0x313111) received by PID 3223825 (TID 0xffff78cb7150) from PID 3223825 ***]

Segmentation fault (core dumped)
root@localhost:/work/env/inference_models/PP-StructrueV3# [ERROR] TBE Subprocess[task_distribute] raise error[], main process disappeared!
[ERROR] TBE Subprocess[task_distribute] raise error[], main process disappeared!
[ERROR] TBE Subprocess[task_distribute] raise error[], main process disappeared!
[ERROR] TBE Subprocess[task_distribute] raise error[], main process disappeared!
[ERROR] TBE Subprocess[task_distribute] raise error[], main process disappeared!
[ERROR] TBE Subprocess[task_distribute] raise error[], main process disappeared!
[ERROR] TBE Subprocess[task_distribute] raise error[], main process disappeared!
[ERROR] TBE Subprocess[task_distribute] raise error[], main process disappeared!
/usr/lib/python3.10/multiprocessing/resource_tracker.py:224: UserWarning: resource_tracker: There appear to be 30 leaked semaphore objects to clean up at shutdown
warnings.warn('resource_tracker: There appear to be %d '
以上是报错信息。初始化分代码如下：
pipeline_npu = PPStructureV3(
device="npu:5",
use_doc_orientation_classify=True,
use_doc_unwarping=True,
use_textline_orientation=True,
use_seal_recognition=True,
use_table_recognition=False,
use_formula_recognition=True,
use_chart_recognition=True,
use_region_detection=True,

# --- 基础模型 ---
layout_detection_model_dir="/work/env/inference_models/PP-StructrueV3/PP-DocLayout_plus-L",
region_detection_model_dir="/work/env/inference_models/PP-StructrueV3/PP-DocBlockLayout",

# --- OCR  ---
text_detection_model_dir="/work/env/inference_models/PP-StructrueV3/PP-OCRv5_server_det",
text_recognition_model_dir="/work/env/inference_models/PP-StructrueV3/PP-OCRv5_server_rec",

# --- 文档/文本方向 (注意参数名) ---
doc_orientation_classify_model_dir="/work/env/inference_models/PP-StructrueV3/PP-LCNet_x1_0_doc_ori",
textline_orientation_model_dir="/work/env/inference_models/PP-StructrueV3/PP-LCNet_x1_0_textline_ori",
table_orientation_classify_model_dir="/work/env/inference_models/PP-StructrueV3/PP-LCNet_x1_0_table_cls",

# --- 通用文档分类 ---
table_classification_model_dir="/work/env/inference_models/PP-StructrueV3/PP-LCNet_x1_0_table_cls",

# --- 文档矫正 ---
doc_unwarping_model_dir="/work/env/inference_models/PP-StructrueV3/UVDoc",

# --- 印章 / 公式 / 图表 ---
seal_text_detection_model_dir = "/work/env/inference_models/PP-StructrueV3/PP-OCRv4_server_seal_det",
seal_text_recognition_model_dir = "/work/env/inference_models/PP-StructrueV3/PP-OCRv5_server_rec",
formula_recognition_model_dir = "/work/env/inference_models/PP-StructrueV3/PP-FormulaNet_plus-L",
chart_recognition_model_dir = "/work/env/inference_models/PP-StructrueV3/PP-Chart2Table"

)

pipeline_cpu = PPStructureV3(

device="cpu",
use_doc_orientation_classify=False,
use_doc_unwarping=False,
use_textline_orientation=False,
use_seal_recognition=False,
use_table_recognition=True,
use_formula_recognition=False,
use_chart_recognition=False,
use_region_detection=False,
# --- 基础模型 ---
layout_detection_model_dir="/work/env/inference_models/PP-StructrueV3/PP-DocLayout_plus-L",
region_detection_model_dir="/work/env/inference_models/PP-StructrueV3/PP-DocBlockLayout",
# --- OCR  ---
text_detection_model_dir="/work/env/inference_models/PP-StructrueV3/PP-OCRv5_server_det",
text_recognition_model_dir="/work/env/inference_models/PP-StructrueV3/PP-OCRv5_server_rec",

# --- 文档/文本方向 (注意参数名) ---
doc_orientation_classify_model_dir="/work/env/inference_models/PP-StructrueV3/PP-LCNet_x1_0_doc_ori",
textline_orientation_model_dir="/work/env/inference_models/PP-StructrueV3/PP-LCNet_x1_0_textline_ori",
table_orientation_classify_model_dir="/work/env/inference_models/PP-StructrueV3/PP-LCNet_x1_0_table_cls",
table_classification_model_dir="/work/env/inference_models/PP-StructrueV3/PP-LCNet_x1_0_table_cls",


chart_recognition_model_dir = "/work/env/inference_models/PP-StructrueV3/PP-Chart2Table",

# --- 表格结构化 ---
wireless_table_cells_detection_model_dir="/work/env/inference_models/PP-StructrueV3/RT-DETR-L_wireless_table_cell_det",
wired_table_cells_detection_model_dir="/work/env/inference_models/PP-StructrueV3/RT-DETR-L_wired_table_cell_det",
wireless_table_structure_recognition_model_dir="/work/env/inference_models/PP-StructrueV3/SLANet_plus",
wired_table_structure_recognition_model_dir="/work/env/inference_models/PP-StructrueV3/SLANeXt_wired"

)

[liuhongen1234567]

您好，由于缺乏相关npu机器，v3中的部分模型可能和您的npu版本不适配。可以将 use_table_recognition=False试试。

[aa851182472]

大佬您好，将 use_table_recognition=False就可以跑通了，但这样做无法实现表格结构解析功能了，是否是因为当前PPStructureV3表格解析模型kernel算子暂不适配昇腾910B2芯片？