Mac M1芯片 Object Detection API 安装

作者: 杨康chin | 来源:发表于2021-09-27 18:12 被阅读0次

Mac M1芯片 Object Detection API 安装
Object Detection API（1）——安装及测试
2022.4.9 mac os M1 芯片 12.3.1 Mon
21个项目玩转深度学习—基于tensorflow的实践详解
基于Tensorflow的自定义对象识别检测模型的训练及视频实时
Python six模块
Cannot install in Homebrew on AR
安装TensorFlow2 Object Detection的s
M1芯片Mac安装homebrew报错及处理
MacOs M1安装Homebrew 在国内最简单方法

十分艰难，环境比较复杂，版本冲突很多，最后终于调试好了。可以直接用这个环境（也包含了tensorflow2.5，要求mac OS 12+）。
tensorflow2.5安装可参考：https://www.jianshu.com/p/0b5342f4af95

先下载 conda环境

wget -c 'https://github.com/chenyangkang/Mac_m1_Object_detection_API_tensorflow25_environment/blob/main/tf_environment/environment.yml'

conda env create -f environment.yml -n tf25
conda activate tf25

再下载pip安装的环境

wget -c 'https://github.com/chenyangkang/Mac_m1_Object_detection_API_tensorflow25_environment/blob/main/tf_environment/requirements.txt'
pip install -r requirements.txt

最后下载测试脚本

wget https://github.com/chenyangkang/Mac_m1_Object_detection_API_tensorflow25_environment/blob/main/tf_environment/model_builder_tf2_test.py
python model_builder_tf2_test.py

输出结果：

Init Plugin
Init Graph Optimizer
Init Kernel
Running tests under Python 3.8.11: /Users/chenyangkang/miniforge3/envs/tf25/bin/python
[ RUN      ] ModelBuilderTF2Test.test_create_center_net_deepmac
Metal device set to: Apple M1

systemMemory: 16.00 GB
maxCacheSize: 5.33 GB

2021-09-27 14:32:24.426421: I tensorflow/core/common_runtime/pluggable_device/pluggable_device_factory.cc:305] Could not identify NUMA node of platform GPU ID 0, defaulting to 0. Your kernel may not have been built with NUMA support.
2021-09-27 14:32:24.426525: I tensorflow/core/common_runtime/pluggable_device/pluggable_device_factory.cc:271] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 0 MB memory) -> physical PluggableDevice (device: 0, name: METAL, pci bus id: <undefined>)
/Users/chenyangkang/miniforge3/envs/tf25/lib/python3.8/site-packages/object_detection/builders/model_builder.py:1091: DeprecationWarning: The 'warn' function is deprecated, use 'warning' instead
  logging.warn(('Building experimental DeepMAC meta-arch.'
W0927 14:32:24.508294 4303684992 model_builder.py:1091] Building experimental DeepMAC meta-arch. Some features may be omitted.
INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_center_net_deepmac): 0.21s
I0927 14:32:24.631808 4303684992 test_util.py:2122] time(__main__.ModelBuilderTF2Test.test_create_center_net_deepmac): 0.21s
[       OK ] ModelBuilderTF2Test.test_create_center_net_deepmac
[ RUN      ] ModelBuilderTF2Test.test_create_center_net_model0 (customize_head_params=True)
INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_center_net_model0 (customize_head_params=True)): 0.25s
I0927 14:32:24.884076 4303684992 test_util.py:2122] time(__main__.ModelBuilderTF2Test.test_create_center_net_model0 (customize_head_params=True)): 0.25s
[       OK ] ModelBuilderTF2Test.test_create_center_net_model0 (customize_head_params=True)
[ RUN      ] ModelBuilderTF2Test.test_create_center_net_model1 (customize_head_params=False)
INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_center_net_model1 (customize_head_params=False)): 0.09s
I0927 14:32:24.975104 4303684992 test_util.py:2122] time(__main__.ModelBuilderTF2Test.test_create_center_net_model1 (customize_head_params=False)): 0.09s
[       OK ] ModelBuilderTF2Test.test_create_center_net_model1 (customize_head_params=False)
Det BiFPN num filters: 112
I0927 14:32:27.176032 4303684992 ssd_efficientnet_bifpn_feature_extractor.py:145] EfficientDet BiFPN num iterations: 5
I0927 14:32:27.176779 4303684992 efficientnet_model.py:147] round_filter input=32 output=32
I0927 14:32:27.183969 4303684992 efficientnet_model.py:147] round_filter input=32 output=32
I0927 14:32:27.184037 4303684992 efficientnet_model.py:147] round_filter input=16 output=16
I0927 14:32:27.275817 4303684992 efficientnet_model.py:147] round_filter input=16 output=16
I0927 14:32:27.275893 4303684992 efficientnet_model.py:147] round_filter input=24 output=24
I0927 14:32:27.350471 4303684992 efficientnet_model.py:147] round_filter input=24 output=24
I0927 14:32:27.350543 4303684992 efficientnet_model.py:147] round_filter input=40 output=48
I0927 14:32:27.427790 4303684992 efficientnet_model.py:147] round_filter input=40 output=48
I0927 14:32:27.427863 4303684992 efficientnet_model.py:147] round_filter input=80 output=88
I0927 14:32:27.536183 4303684992 efficientnet_model.py:147] round_filter input=80 output=88
I0927 14:32:27.536257 4303684992 efficientnet_model.py:147] round_filter input=112 output=120
I0927 14:32:27.645802 4303684992 efficientnet_model.py:147] round_filter input=112 output=120
I0927 14:32:27.645879 4303684992 efficientnet_model.py:147] round_filter input=192 output=208
I0927 14:32:27.791429 4303684992 efficientnet_model.py:147] round_filter input=192 output=208
I0927 14:32:27.791507 4303684992 efficientnet_model.py:147] round_filter input=320 output=352
I0927 14:32:27.860005 4303684992 efficientnet_model.py:147] round_filter input=1280 output=1408
.......
.......
----------------------------------------------------------------------
Ran 24 tests in 9.736s

OK (skipped=1)

运行正常

其他的参考官方流程：
https://tensorflow-object-detection-api-tutorial.readthedocs.io/en/latest/install.html#