108 lines
5.1 KiB
Python
108 lines
5.1 KiB
Python
import random
|
||
import math
|
||
import torch
|
||
import torch.optim as optim
|
||
import torch.nn.functional as F
|
||
from DQN_cnn.memory import ReplayBuffer
|
||
from DQN_cnn.model import CNN
|
||
|
||
|
||
class DQNcnn:
|
||
def __init__(self, screen_height,screen_width, action_dim, cfg):
|
||
|
||
self.device = cfg.device
|
||
self.action_dim = action_dim
|
||
self.gamma = cfg.gamma
|
||
# e-greedy策略相关参数
|
||
self.actions_count = 0
|
||
self.epsilon = 0
|
||
self.epsilon_start = cfg.epsilon_start
|
||
self.epsilon_end = cfg.epsilon_end
|
||
self.epsilon_decay = cfg.epsilon_decay
|
||
self.batch_size = cfg.batch_size
|
||
self.policy_net = CNN(screen_height, screen_width,
|
||
action_dim).to(self.device)
|
||
self.target_net = CNN(screen_height, screen_width,
|
||
action_dim).to(self.device)
|
||
self.target_net.load_state_dict(self.policy_net.state_dict()) # target_net的初始模型参数完全复制policy_net
|
||
self.target_net.eval() # 不启用 BatchNormalization 和 Dropout
|
||
self.optimizer = optim.RMSprop(self.policy_net.parameters(),lr = cfg.lr) # 可查parameters()与state_dict()的区别，前者require_grad=True
|
||
self.loss = 0
|
||
self.memory = ReplayBuffer(cfg.memory_capacity)
|
||
|
||
|
||
def choose_action(self, state):
|
||
'''选择动作
|
||
Args:
|
||
state [array]: [description]
|
||
Returns:
|
||
action [array]: [description]
|
||
'''
|
||
self.epsilon = self.epsilon_end + (self.epsilon_start - self.epsilon_end) * \
|
||
math.exp(-1. * self.actions_count / self.epsilon_decay)
|
||
self.actions_count += 1
|
||
if random.random() > self.epsilon:
|
||
with torch.no_grad():
|
||
q_value = self.policy_net(state) # q_value比如tensor([[-0.2522, 0.3887]])
|
||
# tensor.max(1)返回每行的最大值以及对应的下标，
|
||
# 如torch.return_types.max(values=tensor([10.3587]),indices=tensor([0]))
|
||
# 所以tensor.max(1)[1]返回最大值对应的下标，即action
|
||
action = q_value.max(1)[1].view(1, 1) # 注意这里action是个张量，如tensor([1])
|
||
return action
|
||
else:
|
||
return torch.tensor([[random.randrange(self.action_dim)]], device=self.device, dtype=torch.long)
|
||
|
||
def update(self):
|
||
if len(self.memory) < self.batch_size:
|
||
return
|
||
transitions = self.memory.sample(self.batch_size)
|
||
# Transpose the batch (see https://stackoverflow.com/a/19343/3343043 for
|
||
# detailed explanation). This converts batch-array of Transitions
|
||
# to Transition of batch-arrays.
|
||
batch = self.memory.Transition(*zip(*transitions))
|
||
|
||
# Compute a mask of non-final states and concatenate the batch elements
|
||
# (a final state would've been the one after which simulation ended)
|
||
non_final_mask = torch.tensor(tuple(map(lambda s: s is not None,
|
||
batch.state_)), device=self.device, dtype=torch.bool)
|
||
|
||
non_final_state_s = torch.cat([s for s in batch.state_
|
||
if s is not None])
|
||
state_batch = torch.cat(batch.state)
|
||
action_batch = torch.cat(batch.action)
|
||
reward_batch = torch.cat(batch.reward) # tensor([1., 1.,...,])
|
||
|
||
|
||
# Compute Q(s_t, a) - the model computes Q(s_t), then we select the
|
||
# columns of actions taken. These are the actions which would've been taken
|
||
# for each batch state according to policy_net
|
||
state_action_values = self.policy_net(
|
||
state_batch).gather(1, action_batch) #tensor([[ 1.1217],...,[ 0.8314]])
|
||
|
||
# Compute V(s_{t+1}) for all next states.
|
||
# Expected values of actions for non_final_state_s are computed based
|
||
# on the "older" target_net; selecting their best reward with max(1)[0].
|
||
# This is merged based on the mask, such that we'll have either the expected
|
||
# state value or 0 in case the state was final.
|
||
state__values = torch.zeros(self.batch_size, device=self.device)
|
||
|
||
state__values[non_final_mask] = self.target_net(
|
||
non_final_state_s).max(1)[0].detach()
|
||
|
||
# Compute the expected Q values
|
||
expected_state_action_values = (state__values * self.gamma) + reward_batch # tensor([0.9685, 0.9683,...,])
|
||
|
||
# Compute Huber loss
|
||
self.loss = F.smooth_l1_loss(
|
||
state_action_values, expected_state_action_values.unsqueeze(1)) # .unsqueeze增加一个维度
|
||
# Optimize the model
|
||
self.optimizer.zero_grad() # zero_grad clears old gradients from the last step (otherwise you’d just accumulate the gradients from all loss.backward() calls).
|
||
self.loss.backward() # loss.backward() computes the derivative of the loss w.r.t. the parameters (or anything requiring gradients) using backpropagation.
|
||
for param in self.policy_net.parameters(): # clip防止梯度爆炸
|
||
param.grad.data.clamp_(-1, 1)
|
||
self.optimizer.step() # causes the optimizer to take a step based on the gradients of the parameters.
|
||
|
||
|
||
if __name__ == "__main__":
|
||
dqn = DQN()
|