A Study on the Agent in Fighting Games Based on Deep Reinforcement Learning

Table 2

Winning of rate in training/testing.

Training round	Rate (%)
250	95
Test round
10	100
30	90
50	96