-
Notifications
You must be signed in to change notification settings - Fork 2
/
test_ppo_torch.py
225 lines (184 loc) · 8.79 KB
/
test_ppo_torch.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
222
223
224
225
import time
from collections import deque
import numpy as np
import os
from tqdm import tqdm
import torch
import torch.nn.functional as F
from model_ac_torch import Actor, Critic
import env as env_valid
import fixed_env as env_test
import load_trace
RANDOM_SEED = 42
S_INFO = 6
S_LEN = 8
A_DIM = 6
VIDEO_BIT_RATE = [300,750,1200,1850,2850,4300] # Kbps
BUFFER_NORM_FACTOR = 10.0
CHUNK_TIL_VIDEO_END_CAP = 48.0
M_IN_K = 1000.0
REBUF_PENALTY = 2.66 # 1 sec rebuffering -> 3 Mbps
SMOOTH_PENALTY = 1
DEFAULT_QUALITY = 1
TEST_TRACES_VALID = './cooked_test_traces/'
SUMMARY_DIR = './Results/sim/ppo'
LOG_FILE = './Results/sim/ppo/log'
TEST_LOG_FOLDER = './Results/sim/ppo/test_results/'
LOG_FILE_VALID = './Results/sim/ppo/test_results/log_valid_ppo'
TEST_LOG_FOLDER_VALID = './Results/sim/ppo/test_results/'
# LOG_FILE_TEST = './Results/test/BC/log_hybrid_ppo'
# SUMMARY_DIR = './Results/test/BC/'
Log_path = './Results/sim/ppo'
dtype = torch.cuda.FloatTensor if torch.cuda.is_available() else torch.FloatTensor
dlongtype = torch.cuda.LongTensor if torch.cuda.is_available() else torch.LongTensor
dshorttype = torch.cuda.ShortTensor if torch.cuda.is_available() else torch.ShortTensor
def evaluation(model, log_path_ini, net_env, all_file_name, detail_log = True):
# all_file_name = net_env.get_file_name()
state = np.zeros((S_INFO,S_LEN))
state = torch.from_numpy(state)
# reward_sum = 0
done = True
last_bit_rate = DEFAULT_QUALITY
bit_rate = DEFAULT_QUALITY
# model.load_state_dict(model.state_dict())
log_path = log_path_ini + '_' + all_file_name[net_env.trace_idx]
log_file = open(log_path, 'w')
time_stamp = 0
for video_count in tqdm(range(len(all_file_name))):
while True:
with torch.no_grad():
prob= model(state.unsqueeze(0).type(dtype))
action = prob.multinomial(num_samples=1).detach()
bit_rate = int(action.squeeze().cpu().numpy())
if detail_log == False:
delay, sleep_time, buffer_size, rebuf, \
video_chunk_size, next_video_chunk_sizes, \
end_of_video, video_chunk_remain = \
net_env.get_video_chunk(bit_rate) ## sample in the environment of virtual player
time_stamp += delay # in ms
time_stamp += sleep_time # in ms
# reward is video quality - rebuffer penalty - smooth penalty
# -- lin scale reward --
# reward = VIDEO_BIT_RATE[bit_rate] / M_IN_K \
# - REBUF_PENALTY * rebuf \
# - SMOOTH_PENALTY * np.abs(VIDEO_BIT_RATE[bit_rate] -
# VIDEO_BIT_RATE[last_bit_rate]) / M_IN_K
# -- log scale reward --
log_bit_rate = np.log(VIDEO_BIT_RATE[bit_rate] / float(VIDEO_BIT_RATE[0]))
log_last_bit_rate = np.log(VIDEO_BIT_RATE[last_bit_rate] / float(VIDEO_BIT_RATE[0]))
reward = log_bit_rate \
- REBUF_PENALTY * rebuf \
- SMOOTH_PENALTY * np.abs(log_bit_rate - log_last_bit_rate)
last_bit_rate = bit_rate
log_file.write(str(time_stamp / M_IN_K) + '\t' +
str(VIDEO_BIT_RATE[bit_rate]) + '\t' +
str(buffer_size) + '\t' +
str(rebuf) + '\t' +
str(video_chunk_size) + '\t' +
str(delay) + '\t' +
str(reward) + '\n')
log_file.flush()
# dequeue history record
state = np.roll(state, -1, axis=1)
# this should be S_INFO number of terms
state[0, -1] = VIDEO_BIT_RATE[bit_rate] / float(np.max(VIDEO_BIT_RATE)) # last quality
state[1, -1] = buffer_size / BUFFER_NORM_FACTOR # 10 sec
state[2, -1] = float(video_chunk_size) / float(delay) / M_IN_K # kilo byte / ms
state[3, -1] = float(delay) / M_IN_K / BUFFER_NORM_FACTOR # 10 sec
state[4, :A_DIM] = np.array(next_video_chunk_sizes) / M_IN_K / M_IN_K # mega byte
state[5, -1] = np.minimum(video_chunk_remain, CHUNK_TIL_VIDEO_END_CAP) / float(CHUNK_TIL_VIDEO_END_CAP)
state = torch.from_numpy(state)
if end_of_video:
state = np.zeros((S_INFO,S_LEN))
state = torch.from_numpy(state)
last_bit_rate = DEFAULT_QUALITY
bit_rate = DEFAULT_QUALITY
log_file.write('\n')
log_file.close()
time_stamp = 0
if video_count + 1 >= len(all_file_name):
break
else:
log_path = log_path_ini + '_' + all_file_name[net_env.trace_idx]
log_file = open(log_path, 'w')
break
# else:
# delay, video_chunk_size, video_distortion, reward, next_state, done, rebuf, download_interval = env.step([int(action_1.squeeze()), int(action_2.squeeze()), int(A_DIM-1)]) # Step
# # r_batch.append(reward)
# time_stamp += download_interval
# # log time_stamp, bit_rate, buffer_size, reward
# log_file.write(str(time_stamp) + '\t' +
# str(next_state[0][26]) + '\t' + str(next_state[0][27]) + '\t' + str(next_state[0][28]) + '\t' + # qualities
# str(next_state[0][8]*5.0) + '\t' + ## buffer size
# str(rebuf) + '\t' + ## rebuffering time
# str(reward) + '\t' + # chunk reward
# str(video_distortion) + '\t' +
# str(video_chunk_size) + '\t' +
# str(delay) + '\n')
# log_file.flush()
# a quick hack to prevent the agent from stucking
# actions.append(action[0, 0])
# if actions.count(actions[0]) == actions.maxlen:
# done = True
# if done:
# log_file.write('\n')
# log_file.close()
# state = np.zeros((1,S_LEN))
# time_stamp = 0
# if video_count >= len(all_file_name):
# break
# else:
# log_path = LOG_FILE_VALID + '_' + all_file_name[env.trace_idx]
# log_file = open(log_path, 'w')
# break
# state = torch.from_numpy(state)
def valid(shared_model, epoch, log_file):
os.system('rm -r ' + TEST_LOG_FOLDER_VALID)
os.system('mkdir ' + TEST_LOG_FOLDER_VALID)
model = Actor(A_DIM).type(dtype)
model.eval()
model.load_state_dict(shared_model.state_dict())
log_path_ini = LOG_FILE_VALID
all_cooked_time, all_cooked_bw, all_file_names = load_trace.load_trace(TEST_TRACES_VALID)
env = env_valid.Environment(all_cooked_time=all_cooked_time,
all_cooked_bw=all_cooked_bw)
evaluation(model, log_path_ini, env, all_file_names, False)
rewards = []
test_log_files = os.listdir(TEST_LOG_FOLDER)
for test_log_file in test_log_files:
reward = []
with open(TEST_LOG_FOLDER + test_log_file, 'rb') as f:
for line in f:
parse = line.split()
try:
reward.append(float(parse[-1]))
except IndexError:
break
rewards.append(np.sum(reward[1:]))
rewards = np.array(rewards)
rewards_min = np.min(rewards)
rewards_5per = np.percentile(rewards, 5)
rewards_mean = np.mean(rewards)
rewards_median = np.percentile(rewards, 50)
rewards_95per = np.percentile(rewards, 95)
rewards_max = np.max(rewards)
log_file.write(str(int(epoch)) + '\t' +
str(rewards_min) + '\t' +
str(rewards_5per) + '\t' +
str(rewards_mean) + '\t' +
str(rewards_median) + '\t' +
str(rewards_95per) + '\t' +
str(rewards_max) + '\n')
log_file.flush()
add_str = 'ppo'
model_save_path = Log_path + "/%s_%s_%d.model" %(str('abr'), add_str, int(epoch))
torch.save(shared_model.state_dict(), model_save_path)
def test(test_model, test_traces, log_file):
model = Actor(A_DIM).type(dtype)
model.eval()
model.load_state_dict(torch.load(test_model))
log_path_ini = log_file + 'log_test_ppo'
all_cooked_time, all_cooked_bw, all_file_names = load_trace.load_trace(test_traces)
env = env_test.Environment(all_cooked_time=all_cooked_time,
all_cooked_bw=all_cooked_bw)
evaluation(model, log_path_ini, env, all_file_names, False)