forked from ShayanJa/RL_trader
-
Notifications
You must be signed in to change notification settings - Fork 0
/
learner.py
221 lines (173 loc) · 7.07 KB
/
learner.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
import tensorflow as tf
from tf_agents.agents.dqn import dqn_agent
from tf_agents.networks import q_network
from tf_agents.drivers import dynamic_step_driver
from tf_agents.environments import tf_py_environment
from tf_agents.trajectories import trajectory
from tf_agents.environments import wrappers
from tf_agents.metrics import tf_metrics
from tf_agents.policies import random_tf_policy
from tf_agents.policies import policy_saver
from tf_agents.replay_buffers import tf_uniform_replay_buffer
from tf_agents.utils import common
from tf_agents.metrics import py_metrics
from tf_agents.metrics import tf_metrics
from tf_agents.drivers import py_driver
from tf_agents.drivers import dynamic_episode_driver
from environment import TradingEnvironment
from features import Features
import datetime as dt
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
# Check for GPU
physical_devices = tf.config.experimental.list_physical_devices('GPU')
if physical_devices:
tf.config.experimental.set_memory_growth(physical_devices[0], True)
# Declare env parameters
AGENT_MODEL_PATH = "policy_1min_1000" # @param
SAVE_MODEL, LOAD_MODEL = False, True # @param
STOCK = 'BTC' # @param
CSV_PATH = 'asset_prices/btcusd.csv' # @param
# Declare model Hyper parameters
num_iterations = 10000 # @param
replay_buffer_capacity = 100000 # @param
fc_layer_params = (100,23,33)
batch_size = 100 # @param
learning_rate = 1e-4 # @param
log_interval = 200 # @param
eval_interval = 200 # @param
date_split = dt.datetime(2020, 1, 1, 1, 0) # @param
initial_balance = 100 # @param
training_duration = 1000 # @param
eval_duration = 200 # @param
# Split data into training and test set
prices = pd.read_csv(CSV_PATH, parse_dates=True, index_col=0)
train = prices[:date_split]
test = prices[date_split:]
# Create a feature list:
# Push start date forward by features_length to have non-zero initial features
# Cleanup logic here
features_length = 20
train_features = Features(train, features_length)
test_features = Features(test, features_length)
# Create Environments
train_py_env = wrappers.TimeLimit(TradingEnvironment(initial_balance, train_features), duration=training_duration)
eval_py_env = wrappers.TimeLimit(TradingEnvironment(initial_balance, test_features), duration=eval_duration)
test_py_env = wrappers.TimeLimit(TradingEnvironment(initial_balance, test_features), duration=len(test)-features_length-1)
train_env = tf_py_environment.TFPyEnvironment(train_py_env)
eval_env = tf_py_environment.TFPyEnvironment(eval_py_env)
test_env = tf_py_environment.TFPyEnvironment(test_py_env)
# Initialize Q Network
q_net = q_network.QNetwork(
train_env.observation_spec(),
train_env.action_spec(),
fc_layer_params=fc_layer_params)
optimizer = tf.compat.v1.train.AdamOptimizer(learning_rate=learning_rate)
train_step_counter = tf.compat.v2.Variable(0)
tf_agent = dqn_agent.DqnAgent(
train_env.time_step_spec(),
train_env.action_spec(),
q_network=q_net,
optimizer=optimizer,
train_step_counter=train_step_counter)
tf_agent.initialize()
if LOAD_MODEL:
try:
collect_policy = tf.compat.v2.saved_model.load(AGENT_MODEL_PATH)
policy_state = collect_policy.get_initial_state(batch_size=3)
print("Loading policy from: {}".format(AGENT_MODEL_PATH))
except:
print("Initiating new policy...")
collect_policy = tf_agent.collect_policy
else:
print("Initiating new policy...")
collect_policy = tf_agent.collect_policy
replay_buffer = tf_uniform_replay_buffer.TFUniformReplayBuffer(
data_spec=tf_agent.collect_data_spec,
batch_size=train_env.batch_size,
max_length=replay_buffer_capacity)
replay_observer = [replay_buffer.add_batch]
train_metrics = [
tf_metrics.NumberOfEpisodes(),
tf_metrics.EnvironmentSteps(),
tf_metrics.AverageReturnMetric(),
tf_metrics.AverageEpisodeLengthMetric(),
]
def compute_performance(environment, policy):
time_step = environment.reset()
total_return = 0.0
balance = [time_step.observation[0][0]]
actions = []
while not time_step.is_last():
action_step = policy.action(time_step)
time_step = environment.step(action_step.action)
total_return += time_step.reward
balance.append(time_step.observation[0][0])
actions.append(action_step.action)
# print(time_step.observation)
return total_return[0], balance, actions
def collect_step(environment, policy):
time_step = environment.current_time_step()
action_step = policy.action(time_step)
next_time_step = environment.step(action_step.action)
traj = trajectory.from_transition(time_step, action_step, next_time_step)
# Add trajectory to the replay buffer
replay_buffer.add_batch(traj)
for _ in range(1000):
collect_step(train_env, tf_agent.collect_policy)
dataset = replay_buffer.as_dataset(
num_parallel_calls=3,
sample_batch_size=batch_size,
num_steps=2).prefetch(3)
driver = dynamic_step_driver.DynamicStepDriver(
train_env,
collect_policy,
observers=replay_observer + train_metrics,
num_steps=1)
iterator = iter(dataset)
print("Initial Balance: {}".format(initial_balance))
tf_agent.train = common.function(tf_agent.train)
tf_agent.train_step_counter.assign(0)
final_time_step, policy_state = driver.run()
for i in range(1000):
final_time_step, _ = driver.run(final_time_step, policy_state)
episode_len = []
portfolio_balance = []
for i in range(num_iterations):
final_time_step, _ = driver.run(final_time_step, policy_state)
for _ in range(1):
collect_step(train_env, tf_agent.collect_policy)
experience, _ = next(iterator)
train_loss = tf_agent.train(experience=experience)
step = tf_agent.train_step_counter.numpy()
if step % log_interval == 0:
print('step = {0}: loss = {1}'.format(step, train_loss.loss))
episode_len.append(train_metrics[3].result().numpy())
print('Average episode length: {}'.format(train_metrics[3].result().numpy()))
if step % eval_interval == 0:
reward, portfolio_balance, actions = compute_performance(eval_env, tf_agent.policy)
print('step = {0}: Average Reward = {1}: Ending Portfolio Balance = {2}'.format(step, reward, portfolio_balance[-1]))
if SAVE_MODEL:
print("Saving Model...")
my_policy = tf_agent.collect_policy
saver = policy_saver.PolicySaver(my_policy, batch_size=None)
saver.save('policy_1min_%d' % num_iterations)
# Compare against Buy and hold
reward, portfolio_balance, actions = compute_performance(test_env, tf_agent.policy)
bnh = eval_py_env.buy_and_hold()
portfolio_balance = pd.DataFrame(data={'Close': np.array(portfolio_balance)}, index=bnh.index)
rl_gain = portfolio_balance.iloc[[0,-1]]['Close'].diff().values[-1]
bnh_gain = bnh.iloc[[0,-1]]['Close'].diff().values[-1]
print("RL GAIN = {}: BUY/HOLD = {}".format(rl_gain, bnh_gain))
print("% RL GAIN = {}: % BUY/HOLD = {}".format( rl_gain / portfolio_balance.iloc[0]['Close'], bnh_gain / bnh.iloc[0]['Close']))
bnh['Close'].plot()
portfolio_balance['Close'].plot()
history = pd.DataFrame(data={"Buy/Hold": bnh['Close'], "RL": portfolio_balance['Close']}, index=bnh.index)
history.to_csv("results/trader-{}".format(dt.datetime.now()))
# Display data
plt.title(STOCK)
plt.xlabel('Date')
plt.ylabel('Price')
plt.savefig("results/" + STOCK + ".png")
plt.show()