cuda error: device-side accert triggred

[ops-admin-chengdu]

Description

处理部分字段为空或者含有特殊字符的csv文件时出现异常。

Reproduce

python3 1.py

1.py 的内容

import os
import csv
from opendlp.sensitive_analyze import table_analyzer

import tracemalloc

#tracemalloc.start()


# 定义目标目录
output_directory = 'log-new/result'
# 打开 csv_paths.txt 文件以逐行扫描



with open('csv_paths.txt', 'r') as file:
    for line in file:
        csv_table_path = line.strip()  # 去除行尾的换行符和空白
        print(csv_table_path)

        # 检查文件是否存在
        if os.path.exists(csv_table_path):
            result = table_analyzer.analyze(csv_table_path)

            # 创建一个新的输出文件路径
            output_file_name = os.path.join(output_directory, f"output_analyze_{os.path.basename(csv_table_path)}.txt")
            # 确保目标目录存在，如果不存在则创建
            os.makedirs(os.path.dirname(output_file_name), exist_ok=True)
            with open(output_file_name, 'w', newline='') as output_csvfile:
                writer = csv.writer(output_csvfile, delimiter=',')
                for k, v in result.items():
                    writer.writerow([k, v])
        else:
            print(f"File does not exist: {csv_table_path}")

Expected behavior

Context

• Operating System and version: ubuntu16.04
• Browser and version(if necessary): none
• Which version are you using:
  python3.6.10
  tldextract==3.1.2
  Ipy==1.1
  pandas==1.1.5
  torch==1.10.0
  hanlp==2.1.0a64
  pytest==7.0.1

GPU: tesla P4

Error message

/pytorch/aten/src/ATen/native/cuda/Indexing.cu:699: indexSelectLargeIndex: block: [32,0,0], thread: [94,0,0] Assertion `srcIndex < srcSelectDimSize` failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:699: indexSelectLargeIndex: block: [32,0,0], thread: [95,0,0] Assertion `srcIndex < srcSelectDimSize` failed.
    return self._apply(convert)
  File "/root/anaconda3/envs/python36/lib/python3.6/site-packages/torch/nn/modules/module.py", line 570, in _apply
    module._apply(fn)
  File "/root/anaconda3/envs/python36/lib/python3.6/site-packages/torch/nn/modules/module.py", line 593, in _apply
    param_applied = fn(param)
  File "/root/anaconda3/envs/python36/lib/python3.6/site-packages/torch/nn/modules/module.py", line 897, in convert
    return t.to(device, dtype if t.is_floating_point() or t.is_complex() else None, non_blocking)
RuntimeError: CUDA error: device-side assert triggered

Configuration

csv文件内容举例：
# cat 1.csv （1.csv部分内容为空时，抛出如上错误信息）
id,sample_flow_id,order_id,sale_assistant_id,opt_manager_id,opt_director_id,area_opt_holder_id,audit_person_id,audit_time,audit_valid,invalid_reason,reason_attachment,area_holder_id,sample_priority,sample_total
1300668,58,2048434,,10149824,,,,,,,,,27,147
1300666,58,2048432,,10149824,,,,,,,,,89,147
1300664,58,2048422,,10149610,,,,,,,,,3,5
1300662,58,2048420,,10149824,,,,,,,,,46,147
1300660,58,2048406,,10045256,,,,,,,,,4,4
1300658,58,2048404,,10149824,,,,,,,,,10,147
1300656,58,2048402,,10149824,,,,,,,,,3,147
1300654,58,2048382,,10184292,,,,,,,,,0,1
1300652,58,2048372,,10149610,,,,,,,,,2,2
1300650,58,2048364,,10045256,,,,,,,,,5,11
#cat 2.csv (含有中文)

id,name,vehicle_photo_type_id,terminal_photo_type_id,is_active,create_at,update_at,version

54,其他,9,5,1,2020-11-13 02:46:15,2020-11-13 02:46:15,0

52,照片模糊,9,5,1,2020-11-13 02:46:15,2020-11-13 02:46:15,0

50,其他,9,4,1,2020-11-13 02:46:15,2020-11-13 02:46:15,0

48,照片模糊,9,4,1,2020-11-13 02:46:15,2020-11-13 02:46:15,0

46,设备号/卡号与照片不一致,9,1,1,2020-08-03 23:09:53,2020-08-03 23:09:53,0

44,设备号/卡号照片模糊,9,1,1,2020-08-03 23:09:53,2020-08-03 23:09:53,0

42,安装位置与照片不一致,9,3,1,2020-08-03 23:09:53,2020-08-03 23:09:53,0

40,安装位置照片模糊,9,3,1,2020-08-03 23:09:53,2020-08-03 23:09:53,0

38,车架号与照片不一致,4,,1,2020-08-03 23:09:53,2020-08-03 23:09:53,0

36,车架号照片模糊,4,,1,2020-08-03 23:09:53,2020-08-03 23:09:53,0

Additional context

Add any other context about the problem here.