A long-standing goal of artificial intelligence is an algorithm that learns, tabula rasa, superhuman proficiency in challenging domains. --D.Silver
This is implementation for FizzBuzz using AlphaZero method.
日本語の解説はこちら
- Python 2.7 / 3.5 or later
- TensorFlow 1.4 or later
To learn (about 10min with CPU):
$ python fizzbuzzzero.py --learn
0 total games / next epoch: 1
100/100 games
match: accuracy=25.9[%] average length=0.3
train: accuracy=29.0[%] mse=0.500
100 total games / next epoch: 2
100/100 games
match: accuracy=23.1[%] average length=0.3
train: accuracy=24.3[%] mse=0.499
200 total games / next epoch: 3
(...)
100/100 games
match: accuracy=100.0[%] average length=100.0
train: accuracy=96.2[%] mse=0.172
1900 total games / next epoch: 20
100/100 games
match: accuracy=100.0[%] average length=100.0
accuracy seems to be stable at 100%
Just to test trained network:
$ python fizzbuzzzero.py
<test game>
player lives = (1, 1)
game record ([]=wrong answer):
1, 2, Fizz, 4, Buzz, Fizz, 7, 8, Fizz, Buzz,
11, Fizz, 13, 14, FizzBuzz, 16, 17, Fizz, 19, Buzz,
Fizz, 22, 23, Fizz, Buzz, 26, Fizz, 28, 29, FizzBuzz,
31, 32, Fizz, 34, Buzz, Fizz, 37, 38, Fizz, Buzz,
41, Fizz, 43, 44, FizzBuzz, 46, 47, Fizz, 49, Buzz,
Fizz, 52, 53, Fizz, Buzz, 56, Fizz, 58, 59, FizzBuzz,
61, 62, Fizz, 64, Buzz, Fizz, 67, 68, Fizz, Buzz,
71, Fizz, 73, 74, FizzBuzz, 76, 77, Fizz, 79, Buzz,
Fizz, 82, 83, Fizz, Buzz, 86, Fizz, 88, 89, FizzBuzz,
91, 92, Fizz, 94, Buzz, Fizz, 97, 98, Fizz, Buzz
Other options are as follows.
| option | |
|---|---|
| -h, --help | show help |
| --learn | start to learn |
| --game_cnt GAME_CNT | games to be played in an epoch |
| --serch_limit SEARCH_LIMIT | limit of search count |
| --initial_life INITIAL_LIFE | initial life of each player |
| -v, --verbose | show thinking log of the first game |
| --gpu | enable to use GPU |
| --gpu_cnt GPU_CNT | number of GPUs used for learning |