从object_detection导入model_main失败，TensorFlow无contrib属性报错求助

报错信息 #

Traceback (most recent call last):
  File "../LearningSpace/ObJTEST01/main.py", line 9, in <module>
    from object_detection import model_main
  File "../anaconda3/envs/ObJTESTV01/lib/python3.10/site-packages/object_detection/model_main.py", line 26, in <module>
    from object_detection import model_lib
  File "../anaconda3/envs/ObJTESTV01/lib/python3.10/site-packages/object_detection/model_lib.py", line 27, in <module>
    from object_detection import eval_util
  File "../anaconda3/envs/ObJTESTV01/lib/python3.10/site-packages/object_detection/eval_util.py", line 35, in <module>
    slim = tf.contrib.slim
AttributeError: module 'tensorflow' has no attribute 'contrib'

我的代码 #

import tensorflow as tf
from object_detection.utils import config_util
from object_detection.protos import pipeline_pb2
from google.protobuf import text_format
import os
import requests
import shutil
import tarfile
from object_detection import model_main  # 报错的导入语句

CUSTOM_MODEL_NAME = 'my_ssd_mobnet'
PRETRAINED_MODEL_NAME = 'ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8'
PRETRAINED_MODEL_URL = 'http://download.tensorflow.org/models/object_detection/tf2/20200711/ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8.tar.gz'
TF_RECORD_SCRIPT_NAME = 'generate_tfrecord.py'
LABEL_MAP_NAME = 'label_map.pbtxt'

paths = {
    'WORKSPACE_PATH': os.path.join('Tensorflow', 'workspace'),
    'SCRIPTS_PATH': os.path.join('Tensorflow','scripts'),
    'APIMODEL_PATH': os.path.join('Tensorflow','models'),
    'ANNOTATION_PATH': os.path.join('Tensorflow', 'workspace','annotations'),
    'IMAGE_PATH': os.path.join('Tensorflow', 'workspace','images'),
    'MODEL_PATH': os.path.join('Tensorflow', 'workspace','models'),
    'PRETRAINED_MODEL_PATH': os.path.join('Tensorflow', 'workspace','pre-trained-models'),
    'CHECKPOINT_PATH': os.path.join('Tensorflow', 'workspace','models',CUSTOM_MODEL_NAME),
    'OUTPUT_PATH': os.path.join('Tensorflow', 'workspace','models',CUSTOM_MODEL_NAME, 'export'),
    'TFJS_PATH':os.path.join('Tensorflow', 'workspace','models',CUSTOM_MODEL_NAME, 'tfjsexport'),
    'TFLITE_PATH':os.path.join('Tensorflow', 'workspace','models',CUSTOM_MODEL_NAME, 'tfliteexport'),
    'PROTOC_PATH':os.path.join('Tensorflow','protoc')
 }

files = {
    'PIPELINE_CONFIG':os.path.join('Tensorflow', 'workspace','models', CUSTOM_MODEL_NAME, 'pipeline.config'),
    'TF_RECORD_SCRIPT': os.path.join(paths['SCRIPTS_PATH'], TF_RECORD_SCRIPT_NAME),
    'LABELMAP': os.path.join(paths['ANNOTATION_PATH'], LABEL_MAP_NAME)
}

for path in paths.values():
    if not os.path.exists(path):
        if os.name == 'posix' or os.name == 'nt':
            os.makedirs(path, exist_ok=True)

labels = [{'name':'licence', 'id':1}]

with open(files['LABELMAP'], 'w') as f:
    for label in labels:
        f.write('item { \n')
        f.write('\tname:\'{}\'\n'.format(label['name']))
        f.write('\tid:{}\n'.format(label['id']))
        f.write('}\n')

import subprocess

# 训练命令
command = (
    "--model_dir=Tensorflow/workspace/models/my_ssd_mobnet "
    "--pipeline_config_path=Tensorflow/workspace/models/my_ssd_mobnet/pipeline.config "
    "--num_train_steps=10000"
)

已安装的包 #

pip install tensorflow-macos
pip install tensorflow-object-detection-api

需要用上述命令训练模型，但目前无法正常导入相关包，请求解决。

解决方案 #

1. 核心问题 #

TensorFlow 2.x 已彻底移除tf.contrib模块，你安装的tensorflow-object-detection-api是适配TensorFlow 1.x的版本，导致代码调用tf.contrib.slim时出错。另外，原tf.contrib.slim功能现在作为独立包tensorflow-slim存在。

2. 修复步骤 #

步骤1：卸载冲突包

pip uninstall -y tensorflow-object-detection-api

步骤2：安装TF2兼容依赖

# 确保tensorflow-macos正常安装
pip install tensorflow-macos
# 安装独立的slim库
pip install tensorflow-slim
# 安装TF2官方适配的Object Detection API
pip install tf-models-official

步骤3：调整代码逻辑

• 删除原报错导入语句from object_detection import model_main
• 替换为TF2专用的训练脚本调用方式：

修改代码中的训练部分：

# 指定TF2训练脚本路径（若通过tf-models-official安装，也可直接用`python -m object_detection.model_main_tf2`调用）
TRAINING_SCRIPT = os.path.join(paths['APIMODEL_PATH'], 'research', 'object_detection', 'model_main_tf2.py')
# 拼接完整训练命令
full_command = f"python {TRAINING_SCRIPT} {command}"
# 执行训练
subprocess.run(full_command, shell=True, check=True)

步骤4：验证环境

运行以下代码确认依赖加载正常：

import tensorflow as tf
import tensorflow_slim as slim
from object_detection.utils import config_util
print("Dependencies loaded successfully")

内容的提问来源于stack exchange，提问作者Sandeep Kumar Rudhravaram