Functions — github.com/XR-stb/DQN_WUKONG

CodeHub MCPcopy Create free account

hub / github.com/XR-stb/DQN_WUKONG / functions

Functions227 in github.com/XR-stb/DQN_WUKONG

⨍ Functions 227◇ Types & classes 40

↓ 1 callersFunctionis_window_visible

utils/change_window.py:32

↓ 1 callersMethodjudge

( self, action_name, injured, cur_status, next_status, events,

↓ 1 callersMethodload_config

从 YAML 文件中加载配置

↓ 1 callersFunctionload_data_from_directory

从指定目录加载以指定前缀命名的 CSV 数据文件，并累加每个文件的 Reward 列的总和。

train_data/show_rewards.py:6

↓ 1 callersMethodload_template_once

加载模板图像并缓存，如果已经加载过则直接使用缓存。

↓ 1 callersFunctionmain

↓ 1 callersFunctionmain

utils/find_blood_location.py:9

↓ 1 callersFunctionmain

utils/move_view.py:52

↓ 1 callersFunctionmain

train_data/show_reward_avg.py:70

↓ 1 callersFunctionmain

主函数，生成奖励曲线图和数据文件。

train_data/show_rewards.py:51

↓ 1 callersFunctionmain

bench/grab_bench.py:88

↓ 1 callersFunctionmain_loop

utils/display_game_info.py:150

↓ 1 callersFunctionmove_mouse_to_position

utils/move_view.py:45

↓ 1 callersFunctionmove_window

utils/change_window.py:19

↓ 1 callersFunctionplot_boss_health

绘制 Boss 血量曲线图及每10局的平均血量曲线。

train_data/show_boss_blood.py:5

↓ 1 callersFunctionplot_performance

(win32_times, dxcam_times, num_frames)

bench/grab_bench.py:76

↓ 1 callersFunctionplot_rewards

绘制奖励曲线

train_data/show_reward_avg.py:46

↓ 1 callersFunctionplot_total_rewards

绘制每局结束时的总奖励分曲线和每10局的平均奖励曲线。

train_data/show_rewards.py:25

↓ 1 callersFunctionpreprocess_image

utils/move_view.py:17

↓ 1 callersMethodprocess_color

↓ 1 callersFunctionreload_config

utils/action_tool.py:31

↓ 1 callersMethodreopen_shared_memory

在子进程中重新打开共享内存

↓ 1 callersMethodreset

↓ 1 callersMethodrun

主运行循环

process_handler.py:239

↓ 1 callersMethodsave_boss_health_data

保存 Boss 血量数据到同一个 CSV 文件中。

↓ 1 callersMethodsave_episode_data

将每局的奖励数据保存到 CSV 文件，文件名使用时间戳避免覆盖。

↓ 1 callersMethodsave_model

models/ppo.py:427

↓ 1 callersMethodsave_overall_data

保存所有局的总奖励数据，文件名使用时间戳避免覆盖。

↓ 1 callersMethodsendMouse

(self, dx, dy, buttons)

↓ 1 callersMethodset_offset

(offset_x, offset_y)

↓ 1 callersFunctionset_windows_offset

↓ 1 callersMethodstore_data

(self, s, a, r, s_next, done, logprob_a, this_epo_step)

models/ppo.py:384

↓ 1 callersMethodstore_metadata

将元数据存入共享内存

↓ 1 callersMethodtrain_network

models/ppo.py:299

↓ 1 callersMethodupdate_status

↓ 1 callersMethodupdate_status

更新多个变量的状态。 :param kwargs: 以变量名为键，变量值为值的键值对

utils/display_game_info.py:126

↓ 1 callersMethodupdate_target_network

models/ddqn.py:117

↓ 1 callersMethodupdate_target_network

models/ddqn_resnet.py:115

↓ 1 callersMethodupdate_target_network

models/ppo.py:458

↓ 1 callersMethodupdate_target_network

models/dqn.py:116

↓ 1 callersFunctionwait_for_game_window

(running_event)

↓ 1 callersFunctionwait_for_game_window

utils/display_game_info.py:28

↓ 1 callersMethodwrite_frame_and_status

将画面和状态信息写入共享内存

(self, message, parameter=0)

MethodKeybdInput

(self, code, flags)

(self, frame_buffer_size=3)

初始化动作执行器，并从配置文件加载动作

(self, context, running_event)

process_handler.py:18

(self, train_data_dir)

(self, common = None)

(self, sx, sy, ex, ey)

(self, sx, sy, ex, ey)

(self, sx, sy, ex, ey)

(self, sx, sy, ex, ey, blood_gray_min=117, blood_gray_max=255)

(self, sx, sy, ex, ey, blood_gray_min=117, blood_gray_max=255)

(self, sx, sy, ex, ey, skill_gray_min=90, skill_gray_max=160)

(self, sx, sy, ex, ey)

(self, sx, sy, ex, ey)

(self, sx, sy, ex, ey)

(self, sx, sy, ex, ey)

(self, sx, sy, ex, ey)

(self, sx, sy, ex, ey, hulu_gray_min=120, hulu_gray_max=255)

utils/display_game_info.py:95

(self, state_h, state_w, action_dim, context_dim)

models/ddqn.py:11

(self, config, model_file)

models/base_agent.py:9

(self, state_dim, action_dim, context_dim, config, model_file)

models/ddqn_resnet.py:42

models/ppo.py:14

(self, state_dim, action_dim, net_width)

models/ppo.py:29

(self, state_dim, action_dim, net_width)

models/ppo.py:83

(self, state_dim, net_width)

models/ppo.py:121

(self, state_h, state_w, action_dim, context_dim)

models/dqn.py:12

Method_execute_actions

从队列中执行动作。

Methodchoose_action

models/ddqn.py:60

Methodchoose_action

Choose an action based on the state. Returns: tuple: (action, log_prob)

models/base_agent.py:26

Methodchoose_action

models/ddqn_resnet.py:59

Methodchoose_action

models/dqn.py:61

Functionclasstimeit

装饰器，用于测量函数的执行时间并打印出来。支持类成员函数的装饰。

timing_decorator.py:25

Functionclick_and_crop

(event, x, y, flags, param)

utils/show_game_window.py:32

Methodconv2d_size_out

(size, kernel_size=5, stride=1, padding=2)

models/ddqn.py:18

Methodconv2d_size_out

(size, kernel_size=5, stride=1, padding=2)

models/dqn.py:19

Functioncorrection_window

utils/change_window.py:58

MethoddirectKey

(self, key, direction = None, type = None)

Functiondxcam_grab_screen

bench/grab_bench.py:51

(self, x, context_features)

models/ddqn.py:28

(self, x, context_features)

models/ddqn_resnet.py:33

models/ppo.py:21

(self, state_img, state_vec)

models/ppo.py:69

models/ppo.py:130

(self, x, context_features)

models/dqn.py:29

Methodget_log_prob

(self, state, action)

models/ppo.py:113

Functionget_window_position

utils/change_window.py:10

Functiongrab_screen

grabscreen.py:19

Functioninit_camera

grabscreen.py:10

Methodload_model

models/base_agent.py:52

Functionon_action_finished

utils/action_tool.py:50

Functionon_click

(x, y, button, pressed)

utils/action_tool.py:126

← previous next →101–200 of 227, ranked by callers