-
Notifications
You must be signed in to change notification settings - Fork 85
Expand file tree
/
Copy pathviewer.cpp
More file actions
210 lines (163 loc) · 5.64 KB
/
viewer.cpp
File metadata and controls
210 lines (163 loc) · 5.64 KB
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
#include <madrona/viz/viewer.hpp>
#include "sim.hpp"
#include "mgr.hpp"
#include "types.hpp"
#include <filesystem>
#include <fstream>
#include <optional>
#include <iostream>
using namespace madrona;
using namespace madrona::viz;
static HeapArray<float> readReplayLog(const char *path) {
std::ifstream replay_log(path, std::ios::binary);
replay_log.seekg(0, std::ios::end);
int64_t size = replay_log.tellg();
replay_log.seekg(0, std::ios::beg);
HeapArray<float> log(size / sizeof(float));
replay_log.read((char *)log.data(), (size / sizeof(float)) * sizeof(float));
return log;
}
int main(int argc, char *argv[])
{
using namespace madrona_gpudrive;
constexpr int64_t num_views = 2;
uint32_t num_worlds = 1;
if (argc >= 2) {
num_worlds = (uint32_t)atoi(argv[1]);
}
ExecMode exec_mode = ExecMode::CPU;
if (argc >= 3) {
if (!strcmp("--cpu", argv[2])) {
exec_mode = ExecMode::CPU;
} else if (!strcmp("--cuda", argv[2])) {
exec_mode = ExecMode::CUDA;
}
}
const char *replay_log_path = nullptr;
if (argc >= 4) {
replay_log_path = argv[3];
}
std::optional<HeapArray<float>> replay_log;
uint32_t cur_replay_step = 0;
uint32_t num_replay_steps = 0;
if (replay_log_path != nullptr) {
replay_log = readReplayLog(replay_log_path);
num_replay_steps = replay_log->size() / (num_worlds * num_views * 4);
}
bool enable_batch_renderer =
#ifdef MADRONA_MACOS
false;
#else
true;
#endif
WindowManager wm {};
WindowHandle window = wm.makeWindow("GPUDrive", 1920, 1080);
render::GPUHandle render_gpu = wm.initGPU(0, { window.get() });
Manager mgr({
.execMode = exec_mode,
.gpuID = 0,
.scenes = {"../data/processed/examples/tfrecord-00001-of-01000_307.json"},
.params = {
.polylineReductionThreshold = 1.0,
.observationRadius = 100.0,
.maxNumControlledAgents = 0
},
.enableBatchRenderer = enable_batch_renderer,
.extRenderAPI = wm.gpuAPIManager().backend(),
.extRenderDev = render_gpu.device(),
});
madrona::CountT stepCtr = 0;
// math::Quat initial_camera_rotation = math::Quat::angleAxis(0, math::up).normalize();
math::Quat initial_camera_rotation =
(math::Quat::angleAxis(0, math::up) *
math::Quat::angleAxis(-math::pi / 2.f, math::right)).normalize();
Viewer viewer(mgr.getRenderManager(), window.get(), {
.numWorlds = num_worlds,
.simTickRate = 20,
.cameraMoveSpeed = 20.f,
.cameraPosition = 100.f * math::up,
.cameraRotation = initial_camera_rotation,
});
auto replayStep = [&]() {
if (cur_replay_step == num_replay_steps - 1) {
return true;
}
printf("Step: %u\n", cur_replay_step);
for (uint32_t i = 0; i < num_worlds; i++) {
for (uint32_t j = 0; j < num_views; j++) {
uint32_t base_idx = 0;
base_idx = 4 * (cur_replay_step * num_views * num_worlds +
i * num_views + j);
auto acceleration = (*replay_log)[base_idx];
auto steering = (*replay_log)[base_idx + 1];
auto headAngle = (*replay_log)[base_idx + 2];
printf("%d, %d: %f %f %f\n", i, j, acceleration, steering,
headAngle);
mgr.setAction(i, j, acceleration, steering, headAngle);
}
}
cur_replay_step++;
return false;
};
auto self_printer = mgr.selfObservationTensor().makePrinter();
auto partner_printer = mgr.partnerObservationsTensor().makePrinter();
auto lidar_printer = mgr.lidarTensor().makePrinter();
auto steps_remaining_printer = mgr.stepsRemainingTensor().makePrinter();
auto reward_printer = mgr.rewardTensor().makePrinter();
auto printObs = [&]() {
printf("Self\n");
self_printer.print();
printf("Partner\n");
partner_printer.print();
printf("Lidar\n");
lidar_printer.print();
printf("Steps Remaining\n");
steps_remaining_printer.print();
printf("Reward\n");
reward_printer.print();
printf("\n");
};
viewer.loop(
[&mgr](CountT world_idx, const Viewer::UserInput &input) {
using Key = Viewer::KeyboardKey;
if (input.keyHit(Key::R)) {
mgr.reset({(int)world_idx});
}
(void)world_idx;
},
[&mgr](CountT world_idx, CountT agent_idx,
const Viewer::UserInput &input) {
using Key = Viewer::KeyboardKey;
float steering{0};
const float steeringDelta{math::pi / 8};
float acceleration{0};
const float accelerationDelta{1};
bool shift_pressed = input.keyPressed(Key::Shift);
if (input.keyPressed(Key::W)) {
acceleration += accelerationDelta;
}
if (input.keyPressed(Key::S)) {
acceleration -= accelerationDelta;
}
if (input.keyPressed(Key::D)) {
steering += steeringDelta;
}
if (input.keyPressed(Key::A)) {
steering -= steeringDelta;
}
mgr.setAction(world_idx, agent_idx, acceleration, steering, 0);
}, [&]() {
if (replay_log.has_value()) {
bool replay_finished = replayStep();
if (replay_finished) {
viewer.stopLoop();
}
}
mgr.step();
stepCtr++;
if(stepCtr % 91 == 0) { // TODO: change to episode length
mgr.reset({0});
}
// printObs();
}, []() {});
}