TensorFlow的安裝與CNN測試

0.說明

在Google開源該框架之后便使用真實K40m卡測試，由於生產環境是CentOS6.6的操作系統，但是該框架需要在Python2.7環境下執行，CentOS6.6下折騰了一天沒搞定，后來換成CentOS7，順利跑通

1.系統環境

• python >=2.7
• numpy >=1.9
• gcc >=4.8.2
• cuda 7.0
• java >=1.8
• cudnn 6.5 v2

2.安裝部署

 

1. #安裝依賴，kernel-devel是為了安裝cuda
2. yum -y install gcc python-devel kernel-devel
3. #安裝pip
4. wget --no-check-certificate https://github.com/pypa/pip/archive/1.5.5.tar.gz
5. tar zvxf 1.5.5.tar.gz
6. cd pip-1.5.5/
7. python setup.py install
8. #安裝tensorflow,此時確保服務器可以聯網，會自動下載安裝numpy和six
9. pip install http://dlp.iflytek.com/soft/tensorflow-0.5.0-cp27-none-linux_x86_64.whl

3.測試用例CNN

• 下載訓練數據集

 

1. wget http://dlp.iflytek.com/soft/cifar-10-binary.tar.gz
2. tar -zxvf -C /tmp/cifar10_data

• 執行腳本(默認cpu) 
  cd /root/tensorflow-master/tensorflow/models/image/cifar10 
  python cifar10_train.py 
  

• 采用gpu執行 
  python cifar10_multi_gpu_train.py --num_gpus=4 
  

4.腳本相關說明

 

1. #查看幫助
2. python cifar10_train.py --help
3. --batch_size BATCH_SIZE #一批數據的圖片數量，默認是包含128個examples
4. Number of images to process in a batch.
5. --data_dir DATA_DIR #訓練數據集目錄，默認是/tmp/cifar10_data
6. Path to the CIFAR-10 data directory.
7. --train_dir TRAIN_DIR #訓練目錄
8. Directory where to write event logs and checkpoint.
9. --max_steps MAX_STEPS #最大步數，默認是1000000
10. Number of batches to run.
11. --log_device_placement LOG_DEVICE_PLACEMENT
12. Whether to log device placement.
13. --nolog_device_placement

 

---

Tutorials and Machine Learning Examples — TensorFlow 
http://tensorflow.org/tutorials/deep_cnn/index.md