forked from Linyou/taichi-ngp-renderer
-
Notifications
You must be signed in to change notification settings - Fork 0
/
taichi_ngp.py
1127 lines (932 loc) · 43.7 KB
/
taichi_ngp.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
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
from multiprocessing import shared_memory
import os
import numpy as np
import argparse
from matplotlib import pyplot as plt
from scipy.spatial.transform import Rotation as R
import time
import taichi as ti
from taichi.math import uvec3, vec3, vec2
import wget
import cv2
import platform
from camera import OrbitCamera
from typing import Tuple
def depth2img(depth):
depth = (depth-depth.min())/(depth.max()-depth.min())
depth_img = cv2.applyColorMap((depth*255).astype(np.uint8),
cv2.COLORMAP_TURBO)
return depth_img
arch = ti.cuda if ti._lib.core.with_cuda() else ti.vulkan
# arch = ti.vulkan
if platform.system() == 'Darwin':
block_dim = 64
else:
block_dim = 128
data_type = ti.f16
np_type = np.float16
tf_vec3 = ti.types.vector(3, dtype=data_type)
tf_vec8 = ti.types.vector(8, dtype=data_type)
tf_vec16 = ti.types.vector(16, dtype=data_type)
tf_vec32 = ti.types.vector(32, dtype=data_type)
tf_vec1 = ti.types.vector(1, dtype=data_type)
tf_vec2 = ti.types.vector(2, dtype=data_type)
tf_mat1x3 = ti.types.matrix(1, 3, dtype=data_type)
tf_index_temp = ti.types.vector(8, dtype=ti.i32)
MAX_SAMPLES = 1024
NEAR_DISTANCE = 0.01
SQRT3 = 1.7320508075688772
SQRT3_MAX_SAMPLES = SQRT3/1024
SQRT3_2 = 1.7320508075688772*2
PRETRAINED_MODEL_URL = 'https://github.com/Linyou/taichi-ngp-renderer/releases/download/v0.2/{}.npy'
#<----------------- hash table util code ----------------->
@ti.func
def calc_dt(t, exp_step_factor, grid_size, scale):
return data_type(ti.math.clamp(t*exp_step_factor, SQRT3_MAX_SAMPLES, SQRT3_2*scale/grid_size))
@ti.func
def __expand_bits(v):
v = (v * ti.uint32(0x00010001)) & ti.uint32(0xFF0000FF)
v = (v * ti.uint32(0x00000101)) & ti.uint32(0x0F00F00F)
v = (v * ti.uint32(0x00000011)) & ti.uint32(0xC30C30C3)
v = (v * ti.uint32(0x00000005)) & ti.uint32(0x49249249)
return v
@ti.func
def __morton3D(xyz):
xyz = __expand_bits(xyz)
return xyz[0] | (xyz[1] << 1) | (xyz[2] << 2)
@ti.func
def fast_hash(pos_grid_local):
result = ti.uint32(0)
primes = uvec3(ti.uint32(1), ti.uint32(2654435761), ti.uint32(805459861))
for i in ti.static(range(3)):
result ^= ti.uint32(pos_grid_local[i]) * primes[i]
return result
@ti.func
def under_hash(pos_grid_local, resolution):
result = ti.uint32(0)
stride = ti.uint32(1)
for i in ti.static(range(3)):
result += ti.uint32(pos_grid_local[i] * stride)
stride *= resolution
return result
@ti.func
def grid_pos2hash_index(indicator, pos_grid_local, resolution, map_size):
hash_result = ti.uint32(0)
if indicator == 1:
hash_result = under_hash(pos_grid_local, resolution)
else:
hash_result = fast_hash(pos_grid_local)
return hash_result % map_size
@ti.func
def frexp_bit(x):
exponent = 0
if x != 0.0:
# frac = ti.abs(x)
bits = ti.bit_cast(x, ti.u32)
exponent = ti.i32((bits & ti.u32(0x7f800000)) >> 23) - 127
# exponent = (ti.i32(bits & ti.u32(0x7f800000)) >> 23) - 127
bits &= ti.u32(0x7fffff)
bits |= ti.u32(0x3f800000)
frac = ti.bit_cast(bits, ti.f32)
if frac < 0.5:
exponent -= 1
elif frac > 1.0:
exponent += 1
return exponent
@ti.func
def mip_from_pos(xyz, cascades):
mx = ti.abs(xyz).max()
# _, exponent = _frexp(mx)
exponent = frexp_bit(ti.f32(mx))
# frac, exponent = ti.frexp(ti.f32(mx))
return ti.min(cascades-1, ti.max(0, exponent+1))
@ti.func
def mip_from_dt(dt, grid_size, cascades):
# _, exponent = _frexp(dt*grid_size)
exponent = frexp_bit(ti.f32(dt*grid_size))
# frac, exponent = ti.frexp(ti.f32(dt*grid_size))
return ti.min(cascades-1, ti.max(0, exponent))
#<----------------- hash table util code ----------------->
@ti.func
def random_in_unit_disk():
theta = 2.0 * np.pi * ti.random()
return ti.Vector([ti.sin(theta), ti.cos(theta)])
@ti.func
def random_normal():
x = ti.random() * 2. - 1.
y = ti.random() * 2. - 1.
return tf_vec2(x, y)
@ti.func
def dir_encode_func(dir_):
input = tf_vec32(0.0)
dir = dir_/dir_.norm()
x = dir[0]; y = dir[1]; z = dir[2]
xy= x*y; xz= x*z; yz= y*z; x2= x*x; y2= y*y; z2= z*z
temp = 0.28209479177387814
input[0] = data_type(temp)
input[1] = data_type(-0.48860251190291987*y)
input[2] = data_type(0.48860251190291987*z)
input[3] = data_type(-0.48860251190291987*x)
input[4] = data_type(1.0925484305920792*xy)
input[5] = data_type(-1.0925484305920792*yz)
input[6] = data_type(0.94617469575755997*z2 - 0.31539156525251999)
input[7] = data_type(-1.0925484305920792*xz)
input[8] = data_type(0.54627421529603959*x2 - 0.54627421529603959*y2)
input[9] = data_type(0.59004358992664352*y*(-3.0*x2 + y2))
input[10] = data_type(2.8906114426405538*xy*z)
input[11] = data_type(0.45704579946446572*y*(1.0 - 5.0*z2))
input[12] = data_type(0.3731763325901154*z*(5.0*z2 - 3.0))
input[13] = data_type(0.45704579946446572*x*(1.0 - 5.0*z2))
input[14] = data_type(1.4453057213202769*z*(x2 - y2))
input[15] = data_type(0.59004358992664352*x*(-x2 + 3.0*y2))
return input
@ti.data_oriented
class NGP_fw:
def __init__(self, grid_size, base_res, log2_T, res, level, exp_step_factor):
self.res = res
self.N_rays = res[0] * res[1]
self.grid_size = grid_size
self.grid_size3 = grid_size * grid_size * grid_size
# self.exp_step_factor = exp_step_factor
# self.scale = scale
# rays intersection parameters
# t1, t2 need to be initialized to -1.0
self.hits_t = ti.Vector.field(n=2, dtype=data_type, shape=(self.N_rays))
self.hits_t.fill(-1.0)
# self.center = tf_vec3(0.0, 0.0, 0.0)
self.center = ti.Vector.field(n=3, dtype=ti.f32, shape=())
# self.xyz_min = -tf_vec3(scale, scale, scale)
# self.xyz_max = tf_vec3(scale, scale, scale)
# self.half_size = (self.xyz_max - self.xyz_min) / 2
self.noise_buffer = ti.Vector.field(2, dtype=data_type, shape=(self.N_rays))
self.gen_noise_buffer()
self.rays_o = ti.Vector.field(n=3, dtype=data_type, shape=(self.N_rays))
self.rays_d = ti.Vector.field(n=3, dtype=data_type, shape=(self.N_rays))
# count the number of rays that still alive
self.counter = ti.field(ti.i32, shape=())
self.counter[None] = self.N_rays
# current alive buffer index
self.current_index = ti.field(ti.i32, shape=())
self.current_index[None] = 0
# how many samples that need to run the model
self.model_launch = ti.field(ti.i32, shape=())
# buffer for the alive rays
# why times 2?
self.alive_indices = ti.field(ti.i32, shape=(2*self.N_rays,))
# padd the thread to the factor of block size (thread per block)
self.padd_block_network = ti.field(ti.i32, shape=())
self.padd_block_composite = ti.field(ti.i32, shape=())
# hash table variables
# self.min_samples = 1 if exp_step_factor==0 else 4
# self.per_level_scales = 1.3195079565048218 # hard coded, otherwise it will be have lower percision
self.base_res = base_res
self.max_params = 2**log2_T
self.level = level
# hash table fields
# self.offsets = ti.field(ti.i32, shape=(16,))
# self.hash_map_sizes = ti.field(ti.uint32, shape=(16,))
# self.hash_map_indicator = ti.field(ti.i32, shape=(16,))
self.offsets = ti.types.vector(16, dtype=ti.i32)(0)
self.hash_map_sizes = ti.types.vector(16, dtype=ti.uint32)(0)
self.hash_map_indicator = ti.types.vector(16, dtype=ti.i32)(0)
# buffers that used for points sampling
self.max_samples_per_rays = 1 if exp_step_factor==0. else 4
self.max_samples_shape = self.N_rays * self.max_samples_per_rays
self.xyzs = ti.Vector.field(3, dtype=data_type, shape=(self.max_samples_shape,))
self.dirs = ti.Vector.field(3, dtype=data_type, shape=(self.max_samples_shape,))
self.deltas = ti.field(data_type, shape=(self.max_samples_shape,))
self.ts = ti.field(data_type, shape=(self.max_samples_shape,))
# buffers that store the info of sampled points
self.run_model_ind = ti.field(ti.int32, shape=(self.max_samples_shape,))
self.N_eff_samples = ti.field(ti.int32, shape=(self.N_rays,))
# intermediate buffers for network
self.out_3 = ti.field(data_type, shape=(self.max_samples_shape, 3))
self.out_1 = ti.field(data_type, shape=(self.max_samples_shape,))
self.temp_hit = ti.field(ti.i32, shape=(self.max_samples_shape,))
# results buffers
self.opacity = ti.field(ti.f32, shape=(self.N_rays,))
self.depth = ti.field(ti.f32, shape=(self.N_rays))
self.rgb = ti.Vector.field(3, dtype=ti.f32, shape=(self.N_rays,))
# GUI render buffer (data type must be float32)
self.render_buffer = ti.Vector.field(3, dtype=ti.f32, shape=(res[1], res[0],))
# camera parameters
self.lookat = np.array([0.0, 0.0, -1.0])
self.lookat_change = np.zeros((3,))
self.lookup = np.array([0.0, -1.0, 0.0])
def hash_table_init(self):
print('----GridEncoding----')
print(f'base_resolution: {self.base_res}')
print(f'log_scale:{self.per_level_scales:.5f}')
print(f'feature_per_level: {self.numof_hash_feet} logT: {self.max_params}')
print(f'level: {self.level}')
offset = 0
for i in range(self.level):
resolution = int(np.ceil(self.base_res * np.exp(i*np.log(self.per_level_scales)) - 1.0)) + 1
params_in_level = resolution ** 3
params_in_level = int(resolution ** 3) if params_in_level % 8 == 0 else int((params_in_level + 8 - 1) / 8) * 8
params_in_level = min(self.max_params, params_in_level)
self.offsets[i] = offset
self.hash_map_sizes[i] = params_in_level
self.hash_map_indicator[i] = 1 if resolution ** 3 <= params_in_level else 0
offset += params_in_level
offset *= self.numof_hash_feet
error_mesg = f"hash shape don't match offset: {offset}, hash shape: {self.hash_embedding.shape[0]}, final res: {resolution}"
assert offset == self.hash_embedding.shape[0], error_mesg
def get_direction(self):
w, h = int(self.res[1]), int(self.res[0])
fx = self.K[None][0, 0]
fy = self.K[None][1, 1]
cx, cy = 0.5*w, 0.5*h
x, y = np.meshgrid(
np.arange(w, dtype=np.float32)+ 0.5,
np.arange(h, dtype=np.float32)+ 0.5,
indexing='xy'
)
directions = np.stack([(x-cx)/fx, (y-cy)/fy, np.ones_like(x)], -1)
return directions.reshape(-1, 3)
@ti.kernel
# 根据图像w,h,相机内参(fx,fy,cx,cy)计算出每个pixel的方向(dx,dy,1)
def init_direction(self, w: ti.i32, h: ti.i32):
for i, j in ti.ndrange(w, h):
ori_w = 2*self.K[None][0, 2]
# ori_h = self.K[None][1, 2] * 2,
scale = w / ori_w
# scale_h = h / ori_h
fx = self.cam_fov_scale[None]*self.K[None][0, 0]*scale
fy = self.cam_fov_scale[None]*self.K[None][1, 1]*scale
cx, cy = 0.5*w, 0.5*h
uni_dir = ti.Matrix([[(i+0.5-cx)/fx, (j+0.5-cy)/fy, 1.]], dt=data_type)
self.directions[j*w+i] = uni_dir
def load_model(self, model_path):
print('Loading model from {}'.format(model_path))
model = np.load(model_path, allow_pickle=True).item()
np_hash = model['model.hash_encoder.params'].astype(np_type)
np_sigma = model['model.xyz_encoder.params'].astype(np_type)
np_rgb = model['model.rgb_net.params'].astype(np_type)
np_bitfield = model['model.density_bitfield']
self.rgb_depth = model['model.rgb_depth']
self.cascades = model['model.cascade']
self.numof_hash_feet = 2
self.scale = float(model['model.box_scale'])
self.xyz_min = -tf_vec3(self.scale)
self.xyz_max = tf_vec3(self.scale)
self.xyz_delta = self.xyz_max - self.xyz_min
# self.half_size = self.xyz_delta / 2
self.half_size = ti.Vector.field(3, dtype=ti.f32, shape=())
self.half_size[None] = self.xyz_delta / 2
self.exp_step_factor = 1/256 if self.scale > 0.5 else 0.
self.min_samples = 1 if self.exp_step_factor==0. else 4
self.per_level_scales = model['model.per_level_scale']
self.net_width = model['model.n_neurons']
self.sigma_n_input = model['model.sigma_n_input']
# self.sigma_n_input = 16
self.sigma_n_output = model['model.sigma_n_output']
self.rgb_n_input = model['model.rgb_n_input']
self.rgb_n_output = model['model.rgb_n_output']
# sigma_sm_per_weight = int(self.layer1_base / 128)
# rgb_sm_per_weight = int(self.layer2_base / 128)
self.sigma_model_size = np_sigma.shape[0]
self.rgb_model_size = np_rgb.shape[0]
self.sigma_sm_preload = int(self.sigma_model_size/block_dim)
self.rgb_sm_preload = int(self.rgb_model_size/block_dim)
self.sigma_layer1_base = self.net_width*self.sigma_n_input
self.rgb_layer1_base = self.net_width*self.rgb_n_input
self.rgb_layer2_base = self.rgb_layer1_base+self.net_width*self.net_width
self.hash_embedding= ti.field(dtype=data_type, shape=(np_hash.shape[0],))
self.sigma_weights= ti.field(dtype=data_type, shape=(self.sigma_model_size,))
self.rgb_weights= ti.field(dtype=data_type, shape=(self.rgb_model_size,))
# density_bitfield is used for point sampling
self.density_bitfield = ti.field(ti.uint8, shape=(self.cascades*(self.grid_size**3)//8))
self.hash_embedding.from_numpy(np_hash)
self.sigma_weights.from_numpy(np_sigma)
self.rgb_weights.from_numpy(np_rgb)
self.density_bitfield.from_numpy(np_bitfield)
self.xyzs_embedding = ti.field(data_type, shape=(self.max_samples_shape, self.sigma_n_input))
self.final_embedding = ti.field(data_type, shape=(self.max_samples_shape, self.sigma_n_output))
# use the pre-compute direction and scene pose
self.directions = ti.Matrix.field(n=1, m=3, dtype=data_type, shape=(self.N_rays,))
self.pose = ti.Matrix.field(n=4, m=4, dtype=data_type, shape=())
self.K = ti.Matrix.field(n=3, m=3, dtype=data_type, shape=())
self.cam_fov_scale = ti.field(dtype=data_type, shape=())
self.cam_fov_scale[None] = 1.0
self.K.from_numpy(model['K'].astype(np_type))
self.default_rot = model['poses'][20].astype(np_type)[:3, :3]
self.cam = OrbitCamera(self.default_rot, r=2.5)
# self.pose.from_numpy(model['poses'][20].astype(np_type))
# if self.res[0] != 800 or self.res[1] != 800:
# directions = self.get_direction(model['camera_angle_x'])[:, None, :].astype(np_type)
# else:
# directions = model['directions'][:, None, :].astype(np_type)
# self.directions.from_numpy(directions)
self.init_direction(self.res[1], self.res[0])
print('----model loaded----')
print('K: ')
print(self.K)
print('pose: ')
print(self.pose)
print("half size: ", self.half_size[None])
print('rgb_depth: ', self.rgb_depth)
print('cascades: ', self.cascades)
print('scale: ', self.scale)
print('per_level_scales: ', self.per_level_scales)
print('net_width: ', self.net_width)
print('sigma_n_input: ', self.sigma_n_input)
print('sigma_n_output: ', self.sigma_n_output)
print('rgb_n_input: ', self.rgb_n_input)
print('rgb_n_output: ', self.rgb_n_output)
print('sigma_model_size: ', self.sigma_model_size)
print('rgb_model_size: ', self.rgb_model_size)
print('sigma_sm_preload: ', self.sigma_sm_preload)
print('rgb_sm_preload: ', self.rgb_sm_preload)
@staticmethod
def taichi_init(kernel_profiler):
ti.init(
arch=arch,
offline_cache=True,
kernel_profiler=kernel_profiler,
enable_fallback=False,
)
@staticmethod
def taichi_print_profiler():
ti.profiler.print_kernel_profiler_info()
@ti.kernel
def reset(self):
self.depth.fill(0.0)
self.opacity.fill(0.0)
self.counter[None] = self.N_rays
for i, j in ti.ndrange(self.N_rays, 2):
self.alive_indices[i*2+j] = i
@ti.func
def _ray_aabb_intersec(self, ray_o, ray_d):
inv_d = 1.0 / ray_d
# half_size = tf_vec3(self.half_size[None])
t_min = (self.center[None]-self.half_size[None]-ray_o)*inv_d
t_max = (self.center[None]+self.half_size[None]-ray_o)*inv_d
_t1 = ti.min(t_min, t_max)
_t2 = ti.max(t_min, t_max)
t1 = _t1.max()
t2 = _t2.min()
return tf_vec2(t1, t2)
@ti.kernel
# for each ray generate random noise offset (nx,ny in [-1,1])
def gen_noise_buffer(self):
for i in range(self.N_rays):
self.noise_buffer[i] = random_normal()
# self.noise_buffer[i] = random_in_unit_disk()
# dof: consider len's distortion
# dist_to_focus=1.2 len_dis=0.04
@ti.kernel
def ray_intersect_dof(self, dist_to_focus: float, len_dis: float):
ti.block_local(self.pose)
for i in self.directions:
c2w = self.pose[None]
dir_ori = self.directions[i]
offset = len_dis*self.noise_buffer[i]
offset_m = tf_mat1x3(
[[
offset[0],
offset[1],
0.0,
]]
)
c2w_dir = c2w[:3, :3].transpose()
offset_w = offset_m @ c2w_dir
mat_result = (dir_ori*dist_to_focus) @ c2w_dir - offset_w
ray_d = tf_vec3(mat_result[0, 0], mat_result[0, 1],mat_result[0, 2])
ray_o = c2w[:3, 3] + tf_vec3(offset_w[0, 0], offset_w[0, 1],offset_w[0, 2])
t1t2 = self._ray_aabb_intersec(ray_o, ray_d)
if t1t2[1] > 0.0:
self.hits_t[i][0] = data_type(ti.max(t1t2[0], NEAR_DISTANCE))
self.hits_t[i][1] = t1t2[1]
self.rays_o[i] = ray_o
self.rays_d[i] = ray_d
# for each ray get intersect with aabb box
# self.hits_t: two intersect point t1&t2
# self.rays_o,self.rays_d (ray in world coord)
@ti.kernel
def ray_intersect(self):
ti.block_local(self.pose)
for i in self.directions:
c2w = self.pose[None]
mat_result = self.directions[i] @ c2w[:3, :3].transpose()
ray_d = tf_vec3(mat_result[0, 0], mat_result[0, 1],mat_result[0, 2])
ray_o = c2w[:3, 3]
t1t2 = self._ray_aabb_intersec(ray_o, ray_d)
if t1t2[1] > 0.0:
self.hits_t[i][0] = data_type(ti.max(t1t2[0], NEAR_DISTANCE))
self.hits_t[i][1] = t1t2[1]
self.rays_o[i] = ray_o
self.rays_d[i] = ray_d
@ti.kernel
def raymarching_test_kernel(self, N_samples: int):
self.run_model_ind.fill(0)
for n in ti.ndrange(self.counter[None]):
c_index = self.current_index[None]
r = self.alive_indices[n*2+c_index]
grid_size_inv = 1.0/self.grid_size
ray_o = self.rays_o[r]
ray_d = self.rays_d[r]
t1t2 = self.hits_t[r]
d_inv = 1.0/ray_d
t = t1t2[0]
t2 = t1t2[1]
s = 0
start_idx = n * N_samples
while (t<t2) & (s<N_samples):
# xyz = ray_o + t*ray_d
xyz = ray_o + t*ray_d
# dt is t regular in grid
dt = calc_dt(t, self.exp_step_factor, self.grid_size, self.scale)
mip = ti.max(mip_from_pos(xyz, self.cascades),
mip_from_dt(dt, self.grid_size, self.cascades))
mip_bound = ti.min(ti.pow(2., mip-1), self.scale) # current gird bound
# mip = 0
# mip_bound = 0.5
mip_bound_inv = 1/mip_bound
# nxyz is xyz regular in grid
nxyz = ti.math.clamp(0.5*(xyz*mip_bound_inv+1)*self.grid_size, 0.0, self.grid_size-1.0)
# nxyz = ti.ceil(nxyz)
idx = mip*self.grid_size3 + __morton3D(ti.cast(nxyz, ti.u32))
# occ = density_grid_taichi[idx] > 5.912066756501768
occ = self.density_bitfield[ti.u32(idx//8)] & (1 << ti.u32(idx%8))
if occ: # if occ, get one smaple in ray
sn = start_idx + s
for p in ti.static(range(3)):
self.xyzs[sn][p] = xyz[p]
self.dirs[sn][p] = ray_d[p]
self.run_model_ind[sn] = 1
self.ts[sn] = t
self.deltas[sn] = dt
t += dt
self.hits_t[r][0] = t
s += 1
else: # if not occ, ray marching
txyz = (((nxyz+0.5+0.5*ti.math.sign(ray_d))*grid_size_inv*2-1)*mip_bound-xyz)*d_inv
t_target = t + ti.max(0, txyz.min())
t += calc_dt(t, self.exp_step_factor, self.grid_size, self.scale)
while t < t_target:
t += calc_dt(t, self.exp_step_factor, self.grid_size, self.scale)
self.N_eff_samples[n] = s
if s == 0:
self.alive_indices[n*2+c_index] = -1
@ti.kernel
def rearange_index(self, B: ti.i32):
self.model_launch[None] = 0
for i in ti.ndrange(B):
if self.run_model_ind[i]:
index = ti.atomic_add(self.model_launch[None], 1)
self.temp_hit[index] = i
self.model_launch[None] += 1
self.padd_block_network[None] = ((self.model_launch[None]+ block_dim - 1)// block_dim) * block_dim
# self.padd_block_composite[None] = ((self.counter[None]+ 128 - 1)// 128) *128
@ti.kernel
def hash_encode(self):
# get hash table embedding
# ti.loop_config(block_dim=16)
# for level in ti.static(range(16)):
for sn in ti.ndrange(self.model_launch[None]):
for level in ti.static(range(self.level)):
# normalize to [0, 1], before is [-0.5, 0.5]
# xyz = self.xyzs[self.temp_hit[sn]] + 0.5
xyz = (self.xyzs[self.temp_hit[sn]] - self.xyz_min) / (self.xyz_delta)
offset = self.offsets[level] * self.numof_hash_feet
indicator = self.hash_map_indicator[level]
map_size = self.hash_map_sizes[level]
init_val0 = tf_vec1(0.0)
init_val1 = tf_vec1(1.0)
local_feature_0 = init_val0[0]
local_feature_1 = init_val0[0]
index_temp = tf_index_temp(0)
w_temp = tf_vec8(0.0)
hash_temp_1 = tf_vec8(0.0)
hash_temp_2 = tf_vec8(0.0)
scale = self.base_res * ti.exp(level*ti.log(self.per_level_scales)) - 1.0
resolution = ti.cast(ti.ceil(scale), ti.uint32) + 1
pos = xyz * scale + 0.5
pos_grid_uint = ti.cast(ti.floor(pos), ti.uint32)
pos -= pos_grid_uint
# pos_grid_uint = ti.cast(pos_grid, ti.uint32)
for idx in ti.static(range(8)):
# idx_uint = ti.cast(idx, ti.uint32)
w = init_val1[0]
pos_grid_local = uvec3(0)
for d in ti.static(range(3)):
if (idx & (1 << d)) == 0:
pos_grid_local[d] = pos_grid_uint[d]
w *= data_type(1 - pos[d])
else:
pos_grid_local[d] = pos_grid_uint[d] + 1
w *= data_type(pos[d])
index = ti.int32(grid_pos2hash_index(indicator, pos_grid_local, resolution, map_size))
index_temp[idx] = offset+index*self.numof_hash_feet
w_temp[idx] = w
# local_feature_0 += data_type(w * self.hash_embedding[offset+index*2])
# local_feature_1 += data_type(w * self.hash_embedding[offset+index*2+1])
for idx in ti.static(range(8)):
hash_temp_1[idx] = self.hash_embedding[index_temp[idx]]
hash_temp_2[idx] = self.hash_embedding[index_temp[idx]+1]
for idx in ti.static(range(8)):
local_feature_0 += data_type(w_temp[idx] * hash_temp_1[idx])
local_feature_1 += data_type(w_temp[idx] * hash_temp_2[idx])
self.xyzs_embedding[sn, level*self.numof_hash_feet] = local_feature_0
self.xyzs_embedding[sn, level*self.numof_hash_feet+1] = local_feature_1
@ti.kernel
def sigma_layer(self):
ti.loop_config(block_dim=block_dim)
for sn in ti.ndrange(self.padd_block_network[None]):
tid = sn % block_dim
did_launch_num = self.model_launch[None]
init_val = tf_vec1(0.0)
input_val = ti.simt.block.SharedArray((self.sigma_n_input, block_dim), data_type)
weight = ti.simt.block.SharedArray((self.sigma_model_size,), data_type)
hid1 = ti.simt.block.SharedArray((self.net_width, block_dim), data_type)
hid2 = ti.simt.block.SharedArray((self.net_width, block_dim), data_type)
for i in ti.static(range(self.sigma_sm_preload)):
k = tid*self.sigma_sm_preload+i
weight[k] = self.sigma_weights[k] # mlp weight?
for i in ti.static(range(self.sigma_n_input)):
input_val[i, tid] = self.xyzs_embedding[sn, i] # hash encode output
ti.simt.block.sync()
if sn < did_launch_num:
# input_val = tf_vec32(0.0)
for i in range(self.net_width):
temp = init_val[0]
for j in ti.static(range(self.sigma_n_input)):
temp += input_val[j, tid] * weight[i*self.sigma_n_input+j]
hid1[i, tid] = temp
# ti.simt.block.sync()
for i in range(self.sigma_n_output):
temp = init_val[0]
for j in ti.static(range(self.net_width)):
temp += data_type(ti.max(0.0, hid1[j, tid])) * weight[self.sigma_layer1_base+i*self.net_width+j]
hid2[i, tid] = temp
# ti.simt.block.sync()
self.out_1[self.temp_hit[sn]] = data_type(ti.exp(hid2[0, tid])) # sigma output
for i in ti.static(range(self.sigma_n_output)):
self.final_embedding[sn, i] = hid2[i, tid] # input to rgb mlp
# ti.simt.block.sync()
@ti.kernel
def rgb_layer(self):
ti.loop_config(block_dim=block_dim)
for sn in ti.ndrange(self.padd_block_network[None]):
ray_id = self.temp_hit[sn]
tid = sn % block_dim
did_launch_num = self.model_launch[None]
init_val = tf_vec1(0.0)
weight = ti.simt.block.SharedArray((self.rgb_model_size,), data_type)
hid1 = ti.simt.block.SharedArray((self.net_width, block_dim), data_type)
hid2 = ti.simt.block.SharedArray((self.net_width, block_dim), data_type)
for i in ti.static(range(self.rgb_sm_preload)):
k = tid*self.rgb_sm_preload+i
weight[k] = self.rgb_weights[k]
ti.simt.block.sync()
if sn < did_launch_num:
dir_ = self.dirs[ray_id]
input = dir_encode_func(dir_) # direction encode,what about self.dir_encode()
for i in ti.static(range(16)):
input[16+i] = self.final_embedding[sn, i] # contain position encode,sigma mlp output
for i in range(self.net_width):
temp = init_val[0]
for j in ti.static(range(self.rgb_n_input)):
temp += input[j] * weight[i*self.rgb_n_input+j] # mlp forward
hid1[i, tid] = temp
if ti.static(self.rgb_depth == 2): # two mlp?
for i in range(self.net_width):
temp = init_val[0]
for j in ti.static(range(self.net_width)):
temp += data_type(ti.max(0.0, hid1[j, tid])) * weight[self.rgb_layer1_base+i*self.net_width+j]
hid2[i, tid] = temp
for i in ti.static(range(self.rgb_n_output)):
temp = init_val[0]
for j in ti.static(range(self.net_width)):
temp += data_type(ti.max(0.0, hid2[j, tid])) * weight[self.rgb_layer2_base+i*self.net_width+j]
hid1[i, tid] = temp
for i in range(self.rgb_n_output):
self.out_3[self.temp_hit[sn], i] = data_type(1 / (1 + ti.exp(-hid1[i, tid]))) # rgb output
else:
for i in range(self.rgb_n_output):
temp = init_val[0]
for j in ti.static(range(self.net_width)):
temp += data_type(ti.max(0.0, hid1[j, tid])) * weight[self.net_width*self.rgb_n_input+i*self.net_width+j]
hid2[i, tid] = temp
# ti.simt.block.sync()
for i in ti.static(range(self.rgb_n_output)):
self.out_3[self.temp_hit[sn], i] = data_type(1 / (1 + ti.exp(-hid2[i, tid])))
# ti.simt.block.sync()
@ti.kernel
def composite_test(self, max_samples: ti.i32, T_threshold: data_type):
for n in ti.ndrange(self.counter[None]): # for each ray
N_samples = self.N_eff_samples[n]
if N_samples != 0:
c_index = self.current_index[None]
r = self.alive_indices[n*2+c_index]
T = data_type(1.0 - self.opacity[r])
start_idx = n * max_samples
rgb_temp = tf_vec3(0.0)
depth_temp = tf_vec1(0.0)
opacity_temp = tf_vec1(0.0)
out_3_temp = tf_vec3(0.0)
for s in range(N_samples): # for each sample along ray
sn = start_idx + s
a = data_type(1.0 - ti.exp(-self.out_1[sn]*self.deltas[sn]))
w = a * T # opacity=1,T=0,w=0,in the air without rgb
for i in ti.static(range(3)):
out_3_temp[i] = self.out_3[sn, i]
rgb_temp += w * out_3_temp # sigma*rgb_out
depth_temp[0] += w * self.ts[sn] # sigma*length
opacity_temp[0] += w # occ=sum(sigma)
T *= data_type(1.0 - a)
if T <= T_threshold: # if <0.01, de-alive, stop integral along ray
self.alive_indices[n*2+c_index] = -1
break
self.rgb[r] += rgb_temp
self.depth[r] += depth_temp[0]
self.opacity[r] += opacity_temp[0]
@ti.kernel
def re_order(self, B: ti.i32):
self.counter[None] = 0
c_index = self.current_index[None]
n_index = (c_index + 1) % 2
self.current_index[None] = n_index
for i in ti.ndrange(B):
alive_temp = self.alive_indices[i*2+c_index]
if alive_temp >= 0:
index = ti.atomic_add(self.counter[None], 1)
self.alive_indices[index*2+n_index] = alive_temp
def write_image(self):
rgb_np = self.rgb.to_numpy().reshape(self.res[0], self.res[1], 3)
depth_np = self.depth.to_numpy().reshape(self.res[0], self.res[1])
plt.imsave('taichi_ngp.png', (rgb_np*255).astype(np.uint8))
plt.imsave('taichi_ngp_depth.png', depth2img(depth_np))
def render(self, max_samples, T_threshold, use_dof=False, dist_to_focus=0.8, len_dis=0.0, clear_rgb=True) -> Tuple[float, int, int]:
samples = 0
if clear_rgb:
self.rgb.fill(0.0)
self.reset()
self.gen_noise_buffer()
if use_dof:
self.ray_intersect_dof(dist_to_focus, len_dis)
else:
self.ray_intersect()
while samples < max_samples: # max_sample in one frame
N_alive = self.counter[None]
if N_alive == 0: break
# how many more samples the number of samples add for each ray
N_samples = max(min(self.N_rays//N_alive, 64), self.min_samples)
samples += N_samples
launch_model_total = N_alive * N_samples
# print(f"samples: {samples}, N_alive: {N_alive}, N_samples: {N_samples}")
# raymarching using cascades occ grids, to sampling sufface effective
self.raymarching_test_kernel(N_samples)
self.rearange_index(launch_model_total)
# dir encode for direction
# self.dir_encode()
# hash encode for position(xyz)
self.hash_encode()
# sigma mlp
self.sigma_layer()
# rgb mlp
self.rgb_layer()
# self.FullyFusedMLP()
self.composite_test(N_samples, T_threshold)
self.re_order(N_alive)
return samples, N_alive, N_samples
def render_frame(self, n=1):
t = time.time()
for _ in range(n):
samples, N_alive, N_samples = self.render(max_samples=100, T_threshold=1e-2, clear_rgb=True)
ti.sync()
print(f"samples: {samples}, N_alive: {N_alive}, N_samples: {N_samples}")
print(f'Render time: {1000*(time.time()-t)/n:.2f} ms')
self.write_image()
@ti.kernel
def rgb_to_render_buffer(self, frame: ti.i32):
for i, j in self.render_buffer:
rgb = self.rgb[(self.res[0]-j)*self.res[1]+i]
self.render_buffer[i, j] = rgb / frame
@ti.kernel
def depth_max(self) -> vec2:
max_v = self.depth[0]
min_v = self.depth[0]
for i in ti.ndrange(self.N_rays):
ti.atomic_max(max_v, self.depth[i])
ti.atomic_min(min_v, self.depth[i])
return vec2(max_v, min_v)
@ti.kernel
def depth_to_render_buffer(self, max_min: vec2):
for i, j in self.render_buffer:
max_v = max_min[0]
min_v = max_min[1]
depth = self.depth[(self.res[0]-j)*self.res[1]+i]
pixel = (vec3(depth)-min_v)/(max_v-min_v)
self.render_buffer[i, j] = pixel
def init_cam(self):
self.lookat = self.lookat @ self.pose.to_numpy()[:, :3].T
def render_gui(self):
video_manager = None
# check if the export file exists for snapshot and video
export_dir = './export/'
if not os.path.exists(export_dir):
os.mkdir(export_dir)
H, W = self.res
window = ti.ui.Window('Taichi NGP', (W, H))
canvas = window.get_canvas()
gui = window.get_gui()
last_orbit_x = None
last_orbit_y = None
rotate_speed = 50
movement_speed = 0.03
max_samples_for_rendering = 100
render_time = 0
# white_bg = False
recording = False
show_depth = False
use_dof = False
last_use_dof = False
frame = 0
T_threshold = 1e-2
dist_to_focus = 1.2
len_dis=0.04
self.init_cam()
last_pose = self.pose.to_numpy()
total_frame = 0
last_dist_to_focus = dist_to_focus
last_len_dis = len_dis
box_size_x = self.half_size[None][0]
box_size_y = self.half_size[None][1]
box_size_z = self.half_size[None][2]
last_box_size_x = box_size_x
last_box_size_y = box_size_y
last_box_size_z = box_size_z
center_x = self.center[None][0]
center_y = self.center[None][1]
center_z = self.center[None][2]
last_center_x = center_x
last_center_y = center_y
last_center_z = center_z
cam_fov_scale = 1.0
last_cam_fov_scale = cam_fov_scale
while window.running:
# TODO: make it more efficient
total_frame+=1
if window.is_pressed(ti.ui.RMB):
curr_mouse_x, curr_mouse_y = window.get_cursor_pos()
if last_orbit_x is None or last_orbit_y is None:
last_orbit_x, last_orbit_y = curr_mouse_x, curr_mouse_y
else:
dx = curr_mouse_x - last_orbit_x
dy = curr_mouse_y - last_orbit_y
self.cam.orbit(dx, -dy)
last_orbit_x, last_orbit_y = curr_mouse_x, curr_mouse_y
else:
last_orbit_x = None
last_orbit_y = None
if window.is_pressed('w'):
self.cam.scale(0.2)
if window.is_pressed('s'):
self.cam.scale(-0.2)
if window.is_pressed('a'):
self.cam.pan(100, 0.)
if window.is_pressed('d'):
self.cam.pan(-100, 0.)
if window.is_pressed('e'):
self.cam.pan(0., -100)
if window.is_pressed('q'):
self.cam.pan(0., 100)
if self.cam.params_changed:
self.rgb.fill(0.0)
total_frame = 1
self.cam.params_changed = False
self.pose.from_numpy(self.cam.pose.astype(np_type))
with gui.sub_window("Options", 0.05, 0.05, 0.68, 0.3) as w:
w.text(f'General')
box_size_x = w.slider_float('box size_x', box_size_x, 0.01, 64.0)
box_size_y = w.slider_float('box size_y', box_size_y, 0.01, 64.0)
box_size_z = w.slider_float('box size_z', box_size_z, 0.01, 64.0)
center_x = w.slider_float('center_x', center_x, -64.0, 64.0)
center_y = w.slider_float('center_y', center_y, -64.0, 64.0)
center_z = w.slider_float('center_z', center_z, -64.0, 64.0)
cam_fov_scale = w.slider_float('camera fov', cam_fov_scale, 0.01, 2.0)
T_threshold = w.slider_float('transparency threshold', T_threshold, 0., 1.)
max_samples_for_rendering = w.slider_float("max samples", max_samples_for_rendering, 1, 100)
show_depth = w.checkbox("show depth", show_depth)
# white_bg = w.checkbox("white background", white_bg)
if last_box_size_x != box_size_x or last_box_size_y != box_size_y or last_box_size_z != box_size_z:
last_box_size_x = box_size_x
last_box_size_y = box_size_y
last_box_size_z = box_size_z
self.half_size[None] = vec3(box_size_x, box_size_y, box_size_z)
self.rgb.fill(0.0)
total_frame = 1