-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathscenario_room_change.py
182 lines (151 loc) · 6.4 KB
/
scenario_room_change.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
import random
from typing import List
import torch
from torchvision.transforms import Compose, Normalize, Resize, ToTensor
from PIL import Image
from env.tasks import HomeServiceTaskSampler, HomeServiceTaskType
from experiments.home_service_base import HomeServiceBaseExperimentConfig
from room_change.utils import set_model
from room_change.main import parse_args as rc_parse_args
task_sampler_params = HomeServiceBaseExperimentConfig.stagewise_task_sampler_args(
stage="train", process_ind=0, total_processes=1, headless=False
)
task_sampler: HomeServiceTaskSampler = HomeServiceBaseExperimentConfig.make_sampler_fn(
**task_sampler_params,
task_type=HomeServiceTaskType.SIMPLE_PICK_AND_PLACE,
force_cache_reset=True,
epochs=1,
)
device = torch.device('cuda')
num_tasks = 500
success = 0
home_dicts = {f"Home{k}": [f"FloorPlan{k + i * 100}" for i in [0, 2, 3, 4]] for k in range(1, 31)}
room_ind_to_type = {
0: "Kitchen",
1: "Living Room",
2: "Bedroom",
3: "Bathroom",
}
room_ind_to_label = {
0: "Kitchen",
2: "Living Room",
3: "Bedroom",
4: "Bathroom",
}
command_to_action_str = {
1: "MoveAhead",
2: "RotateRight",
3: "RotateLeft",
4: "ResetScene"
}
rc_opt = rc_parse_args()
rc_opt.pretrained = './room_change/save/models/202110042117-scale30-noaugmentation/best_model.pth'
rc_model = set_model(rc_opt, device)
rc_model.eval()
rc_transform = Compose(
[
Resize(224),
ToTensor(),
Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
]
)
for i in range(num_tasks):
print(f'{i}-th task')
task = task_sampler.next_task(pickup_target="Fork", place_target="DiningTable")
home_number = int(task.env.scene[9:]) % 100
task_home = home_dicts[f"Home{home_number}"]
scene_b = random.choice(task_home)
task.env.controller.reset(scene=scene_b, width=512, height=512, quality="Ultra")
print(f'scene: {scene_b}, label: {room_ind_to_label[int(scene_b[9:]) // 100]}')
t_frames = None
input_command = int(input("input_command: "))
step = 0
suffix = "b"
detection = ""
while input_command != -1:
action_str = command_to_action_str[input_command]
frame = task.env.controller.last_event.frame
img = Image.fromarray(frame)
img.save(f'./test/rc/{step}_{suffix}{detection}.png')
t_frame = rc_transform(img)
if t_frames is None:
t_frames = t_frame.unsqueeze(0)
else:
if t_frames.shape[0] == 16:
t_frames[:-1] = t_frames[1:].clone()
t_frames[-1:] = t_frame.unsqueeze(0).clone()
scene_b_pred = room_ind_to_type[rc_model(t_frames[:8].to(device))[1].sum(dim=0).argmax().cpu().item()]
scene_a_pred = room_ind_to_type[rc_model(t_frames[8:].to(device))[1].sum(dim=0).argmax().cpu().item()]
print(f'pred_b: {scene_b_pred}, pred_a: {scene_a_pred}')
if scene_b_pred != scene_a_pred:
print("scene change detected")
detection = "_detected"
else:
detection = ""
elif t_frames.shape[0] < 16:
t_frames = torch.cat([t_frames, t_frame.unsqueeze(0)], dim=0)
else:
raise RuntimeError()
if input_command in [1, 2, 3]:
task.env.controller.step(action_str)
elif input_command == 4:
scene_a = random.choice(task_home)
task.env.controller.reset(scene=scene_a, width=512, height=512, quality="Ultra")
print("scene changed")
print(f'scene: {scene_a}, label: {room_ind_to_label[int(scene_a[9:]) // 100]}')
if scene_a != scene_b:
suffix = "a"
step += 1
input_command = int(input("input_command: "))
# scene_b = random.choice(task_home)
# task.env.controller.reset(scene=scene_b, width=512, height=512, quality="Ultra")
# t_frames_b = None
# for i in range(8):
# frame = task.env.controller.last_event.frame
# img = Image.fromarray(frame)
# img.save(f'./test/rc/b_{i}.png')
# t_frame = rc_transform(img)
# if t_frames_b is None:
# t_frames_b = t_frame.unsqueeze(0)
# else:
# t_frames_b = torch.cat([t_frames_b, t_frame.unsqueeze(0)], dim=0)
# # task.env.controller.step("RotateRight", degrees=45)
# input_command = int(input("input_command: "))
# action_str = command_to_action_str[input_command]
# # task.env.controller.step(action_str, degrees=45)
# task.env.controller.step(action_str)
# scene_b_pred = room_ind_to_type[rc_model(t_frames_b.to(device))[1].sum(dim=0).argmax().cpu().item()]
# scene_a = random.choice(task_home)
# task.env.controller.reset(scene=scene_a, width=512, height=512, quality="Ultra")
# t_frames_a = None
# for i in range(8):
# frame = task.env.controller.last_event.frame
# img = Image.fromarray(frame)
# img.save(f'./test/rc/a_{i}.png')
# t_frame = rc_transform(img)
# if t_frames_a is None:
# t_frames_a = t_frame.unsqueeze(0)
# else:
# t_frames_a = torch.cat([t_frames_a, t_frame.unsqueeze(0)], dim=0)
# # task.env.controller.step("RotateRight", degrees=45)
# input_command = int(input("input_command: "))
# action_str = command_to_action_str[input_command]
# # task.env.controller.step(action_str, degrees=45)
# task.env.controller.step(action_str)
# scene_a_pred = room_ind_to_type[rc_model(t_frames_a.to(device))[1].sum(dim=0).argmax().cpu().item()]
# print(f'[before] scene: {scene_b}, pred: {scene_b_pred}, label: {room_ind_to_label[int(scene_b[9:]) // 100]}')
# print(f'[after] scene: {scene_a}, pred: {scene_a_pred}, label: {room_ind_to_label[int(scene_a[9:]) // 100]}')
import pdb; pdb.set_trace()
print(f' task: {task.env.current_task_spec.task_type}')
while not task.is_done():
obs = task.get_observations()
# action_ind = int(input(f"action_ind="))
# import pdb; pdb.set_trace()
# action_ind, _ = task.query_expert()
step_result = task.step(action=action_ind)
if step_result.info['action_name'] == "done":
success += 1
print(f'{i}-th task done')
task_sampler.close()
print(f'finishied {num_tasks} tasks')
print(f'Success {success} out of {num_tasks}')