tensorflow mnist 示例运行结果

来源：互联网发布：室内设计什么软件好编辑：程序博客网时间：2024/09/21 09:23

[jiangzhijun@localhost mnist]$ ls -al
总用量 52
drwxrwxr-x. 3 jiangzhijun jiangzhijun 4096 3月 19 20:31 .
drwxrwxr-x. 5 jiangzhijun jiangzhijun 67 5月 18 2016 ..
-rw-rw-r--. 1 jiangzhijun jiangzhijun 2183 5月 18 2016 BUILD
drwxr-xr-x. 2 jiangzhijun jiangzhijun 4096 3月 19 20:37 data
-rw-rw-r--. 1 jiangzhijun jiangzhijun 8632 5月 18 2016 fully_connected_feed.py
-rw-rw-r--. 1 jiangzhijun jiangzhijun 967 5月 18 2016 __init__.py
-rw-rw-r--. 1 jiangzhijun jiangzhijun 1095 5月 18 2016 input_data.py
-rw-rw-r--. 1 jiangzhijun jiangzhijun 5286 5月 18 2016 mnist.py
-rw-rw-r--. 1 jiangzhijun jiangzhijun 2022 5月 18 2016 mnist_softmax.py
-rw-rw-r--. 1 jiangzhijun jiangzhijun 6831 5月 18 2016 mnist_with_summaries.py
[jiangzhijun@localhost mnist]$ python fully_connected_feed.py
Extracting data/train-images-idx3-ubyte.gz
Extracting data/train-labels-idx1-ubyte.gz
Extracting data/t10k-images-idx3-ubyte.gz
Extracting data/t10k-labels-idx1-ubyte.gz
Step 0: loss = 2.28 (0.009 sec)
Step 100: loss = 2.14 (0.004 sec)
Step 200: loss = 1.84 (0.004 sec)
Step 300: loss = 1.54 (0.004 sec)
Step 400: loss = 1.20 (0.004 sec)
Step 500: loss = 0.86 (0.004 sec)
Step 600: loss = 0.84 (0.004 sec)
Step 700: loss = 0.66 (0.004 sec)
Step 800: loss = 0.57 (0.004 sec)
Step 900: loss = 0.51 (0.004 sec)
Training Data Eval:
Num examples: 55000 Num correct: 47726 Precision @ 1: 0.8677
Validation Data Eval:
Num examples: 5000 Num correct: 4356 Precision @ 1: 0.8712
Test Data Eval:
Num examples: 10000 Num correct: 8742 Precision @ 1: 0.8742
Step 1000: loss = 0.56 (0.005 sec)
Step 1100: loss = 0.42 (0.089 sec)
Step 1200: loss = 0.62 (0.004 sec)
Step 1300: loss = 0.43 (0.004 sec)
Step 1400: loss = 0.41 (0.004 sec)
Step 1500: loss = 0.37 (0.004 sec)
Step 1600: loss = 0.29 (0.004 sec)
Step 1700: loss = 0.56 (0.004 sec)
Step 1800: loss = 0.31 (0.004 sec)
Step 1900: loss = 0.24 (0.004 sec)
Training Data Eval:
Num examples: 55000 Num correct: 49258 Precision @ 1: 0.8956
Validation Data Eval:
Num examples: 5000 Num correct: 4513 Precision @ 1: 0.9026
Test Data Eval:
Num examples: 10000 Num correct: 8975 Precision @ 1: 0.8975

0 0