-
Notifications
You must be signed in to change notification settings - Fork 3
/
train.py
413 lines (361 loc) · 20.1 KB
/
train.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
#
# Copyright (C) 2023, Inria
# GRAPHDECO research group, https://team.inria.fr/graphdeco
# All rights reserved.
#
# This software is free for non-commercial, research and evaluation use
# under the terms of the LICENSE.md file.
#
# For inquiries contact [email protected]
#
import os
import torch
from torchmetrics.image import PeakSignalNoiseRatio
from random import randint, seed
from utils.loss_utils import l1_loss, ssim
from gaussian_renderer import render, network_gui, render_through
import sys
from scene import Scene, GaussianModel, BilateralFilter
from utils.general_utils import safe_state
import uuid
import torch.nn.functional as F
import wandb
from tqdm import tqdm
from torch import nn
from lpipsPyTorch import lpips
from utils.image_utils import psnr
from argparse import ArgumentParser, Namespace
from arguments import ModelParams, PipelineParams, OptimizationParams
from torch.utils.tensorboard import SummaryWriter
TENSORBOARD_FOUND = True
from utils.depth_utils import depths_to_points, depth_to_normal
import kornia
def culling(xyz, cams, expansion=2):
cam_centers = torch.stack([c.camera_center for c in cams], 0).to(xyz.device)
span_x = cam_centers[:, 0].max() - cam_centers[:, 0].min()
span_y = cam_centers[:, 1].max() - cam_centers[:, 1].min() # smallest span
span_z = cam_centers[:, 2].max() - cam_centers[:, 2].min()
scene_center = cam_centers.mean(0)
span_x = span_x * expansion
span_y = span_y * expansion
span_z = span_z * expansion
x_min = scene_center[0] - span_x / 2
x_max = scene_center[0] + span_x / 2
y_min = scene_center[1] - span_y / 2
y_max = scene_center[1] + span_y / 2
z_min = scene_center[2] - span_x / 2
z_max = scene_center[2] + span_x / 2
valid_mask = (xyz[:, 0] > x_min) & (xyz[:, 0] < x_max) & \
(xyz[:, 1] > y_min) & (xyz[:, 1] < y_max) & \
(xyz[:, 2] > z_min) & (xyz[:, 2] < z_max)
# print(f'scene mask ratio {valid_mask.sum().item() / valid_mask.shape[0]}')
return valid_mask, scene_center
def prune_low_contribution_gaussians(gaussians, cameras, pipe, bg, K=5, prune_ratio=0.1):
top_list = [None, ] * K
for i, cam in enumerate(cameras):
trans = render(cam, gaussians, pipe, bg, record_transmittance=True)
if top_list[0] is not None:
m = trans > top_list[0]
if m.any():
for i in range(K - 1):
top_list[K - 1 - i][m] = top_list[K - 2 - i][m]
top_list[0][m] = trans[m]
else:
top_list = [trans.clone() for _ in range(K)]
contribution = torch.stack(top_list, dim=-1).mean(-1)
tile = torch.quantile(contribution, prune_ratio)
prune_mask = contribution < tile
gaussians.prune_points(prune_mask)
torch.cuda.empty_cache()
def L1_loss_appearance(image, gt_image, gaussians, view_idx, return_transformed_image=False):
appearance_embedding = gaussians.get_apperance_embedding(view_idx)
# center crop the image
origH, origW = image.shape[1:]
H = origH // 32 * 32
W = origW // 32 * 32
left = origW // 2 - W // 2
top = origH // 2 - H // 2
crop_image = image[:, top:top+H, left:left+W]
crop_gt_image = gt_image[:, top:top+H, left:left+W]
# down sample the image
crop_image_down = torch.nn.functional.interpolate(crop_image[None], size=(H//32, W//32), mode="bilinear", align_corners=True)[0]
crop_image_down = torch.cat([crop_image_down, appearance_embedding[None].repeat(H//32, W//32, 1).permute(2, 0, 1)], dim=0)[None]
mapping_image = gaussians.appearance_network(crop_image_down)
transformed_image = mapping_image * crop_image
if not return_transformed_image:
return l1_loss(transformed_image, crop_gt_image)
else:
transformed_image = torch.nn.functional.interpolate(transformed_image, size=(origH, origW), mode="bilinear", align_corners=True)[0]
return transformed_image
def training(dataset, opt, pipe, testing_iterations, saving_iterations, checkpoint_iterations, checkpoint, debug_from, scale_regu, lambda_opa, mask_normal):
print("scal:", scale_regu, "ld_opa:", lambda_opa)
first_iter = 0
tb_writer = prepare_output_and_logger(dataset)
gaussians = GaussianModel(dataset.sh_degree)
gaussians.bi_filter = BilateralFilter(d=opt.fil_width, sigmaColor=50, sigmaSpace=50)
gaussians.bi_filter.to(device="cuda")
scene = Scene(dataset, gaussians)
gaussians.training_setup(opt)
if checkpoint:
(model_params, first_iter) = torch.load(checkpoint)
gaussians.restore(model_params, opt)
if pipe.no_spike:
gaussians.Vth_opa = nn.Parameter(torch.tensor([0.005]).to(device="cuda").requires_grad_(False))
if pipe.no_cut:
gaussians.Vth_pdf.requires_grad_(False)
gaussians.Vth_pdf *= 0
bg_color = [1, 1, 1] if dataset.white_background else [0, 0, 0]
background = torch.tensor(bg_color, dtype=torch.float32, device="cuda")
iter_start = torch.cuda.Event(enable_timing = True)
iter_end = torch.cuda.Event(enable_timing = True)
viewpoint_stack = None
ema_loss_for_log = 0.0
progress_bar = tqdm(range(first_iter, opt.iterations), desc="Training progress")
first_iter += 1
all_cameras = scene.getTrainCameras().copy()
for idx, camera in enumerate(scene.getTrainCameras()):
camera.idx = idx
for iteration in range(first_iter, opt.iterations + 1):
if network_gui.conn == None:
network_gui.try_connect()
while network_gui.conn != None:
try:
net_image_bytes = None
custom_cam, do_training, pipe.convert_SHs_python, pipe.compute_cov3D_python, keep_alive, scaling_modifer = network_gui.receive()
if custom_cam != None:
net_image = render(custom_cam, gaussians, pipe, background, scaling_modifer)["render"]
net_image_bytes = memoryview((torch.clamp(net_image, min=0, max=1.0) * 255).byte().permute(1, 2, 0).contiguous().cpu().numpy())
network_gui.send(net_image_bytes, dataset.source_path)
if do_training and ((iteration < int(opt.iterations)) or not keep_alive):
break
except Exception as e:
network_gui.conn = None
iter_start.record()
gaussians.update_learning_rate(iteration, opt.opacity_reset_interval, pipe)
# Every 1000 its we increase the levels of SH up to a maximum degree
if iteration % 1000 == 0:
gaussians.oneupSHdegree()
# Pick a random Camera
if not viewpoint_stack:
viewpoint_stack = scene.getTrainCameras().copy()
viewpoint_cam = viewpoint_stack.pop(randint(0, len(viewpoint_stack)-1))
view_mask = viewpoint_cam.image_mask.squeeze(0).to(device="cuda")
# Render
if (iteration - 1) == debug_from:
pipe.debug = True
render_pkg = render(viewpoint_cam, gaussians, pipe, background)
image, viewspace_point_tensor, visibility_filter, radii = render_pkg["render"], render_pkg["viewspace_points"], render_pkg["visibility_filter"], render_pkg["radii"]
# Loss
gt_image = viewpoint_cam.original_image.cuda()
Ll1 = 0.
# use L1 loss for the transformed image if using decoupled appearance
if dataset.use_decoupled_appearance:
Ll1 = L1_loss_appearance(image, gt_image, gaussians, viewpoint_cam.idx)
else:
Ll1 = l1_loss(image, gt_image)
lv_opa = 1 / (gaussians.Vth_opa + 1e-6)
lv_pdf = 1 / (gaussians.get_Vth_pdf.mean() + 1e-6)
max_values, _ = torch.max(gaussians.get_scaling[: ,:2], dim=1)
scale_reg = scale_regu
max_values = max_values[max_values > scale_reg]
lscale = torch.sum(max_values)
ssim_loss = ssim(image, gt_image)
ld_opa = opt.ld_opa
ld_pdf = opt.ld_pdf
ld_scale = 0.0005
lambda_normal = opt.lambda_normal if iteration > 7000 else 0.0
lambda_dist = opt.lambda_dist if iteration > 3000 else 0.0
lambda_smooth = opt.lambda_smooth if iteration > -1 else 0.0
lambda_tv_normal = 0.0 if (iteration > 0 and iteration <= 2000) else 0.0
lambda_tv_depth = opt.lambda_tv_d if (iteration <= 4000) else 0.0
rend_dist = render_pkg["rend_dist"]
rend_normal = render_pkg['rend_normal']
surf_normal = render_pkg['surf_normal']
surf_depth = render_pkg['surf_depth']
depth_map = render_pkg["surf_depth"][0]
if opt.depth_grad_thresh > 0:
depth_map_for_grad = depth_map[None, None]
sobel_kernel_x = torch.tensor([[-1, 0, 1], [-2, 0, 2], [-1, 0, 1]], dtype=depth_map.dtype, device=depth_map.device).view(1, 1, 3, 3)
sobel_kernel_y = torch.tensor([[-1, -2, -1], [0, 0, 0], [1, 2, 1]], dtype=depth_map.dtype, device=depth_map.device).view(1, 1, 3, 3)
depth_map_for_grad = F.pad(depth_map_for_grad, pad=(1, 1, 1, 1), mode="replicate")
depth_grad_x = F.conv2d(depth_map_for_grad, sobel_kernel_x) / 3
depth_grad_y = F.conv2d(depth_map_for_grad, sobel_kernel_y) / 3
depth_grad_mag = torch.sqrt(depth_grad_x ** 2 + depth_grad_y ** 2)
depth_grad_weight = (depth_grad_mag < opt.depth_grad_thresh).float()
depth_grad_mask_dilation = 1
mask_di = depth_grad_mask_dilation
depth_grad_weight = -1 * F.max_pool2d(-1 * depth_grad_weight, mask_di * 2 + 1, stride=1, padding=mask_di)
depth_grad_weight = depth_grad_weight.squeeze().detach()
normal_error = (1 - (rend_normal * surf_normal * view_mask.unsqueeze(0)).sum(dim=0)) * depth_grad_weight
normal_error = normal_error[None]
else:
if (1 - mask_normal):
normal_error = (1 - (rend_normal * surf_normal).sum(dim=0))[None]
else:
normal_error = (1 - (rend_normal * surf_normal * view_mask.unsqueeze(0)).sum(dim=0))[None]
normal_loss = lambda_normal * (normal_error).mean()
dist_loss = lambda_dist * (rend_dist).mean()
_, h, w = gt_image.shape
smooth_loss = kornia.losses.inverse_depth_smoothness_loss(surf_depth.unsqueeze(0), gt_image.unsqueeze(0))
smooth_loss = lambda_smooth * smooth_loss
tv_loss_depth = 0.
tv_loss_normal = 0.
tv_loss_depth += gaussians.TVLoss(surf_depth.permute(1, 2, 0))
tv_loss_normal += gaussians.EdgeAwareTVLoss(rend_normal.permute(1, 2, 0), gt_image.permute(1, 2, 0))
tv_loss = lambda_tv_depth * tv_loss_depth + lambda_tv_normal * tv_loss_normal
if pipe.no_spike:
loss = (1.0 - opt.lambda_dssim) * Ll1 + opt.lambda_dssim * (1.0 - ssim_loss) + ld_pdf * lv_pdf + ld_scale * lscale
elif pipe.no_cut:
loss = (1.0 - opt.lambda_dssim) * Ll1 + opt.lambda_dssim * (1.0 - ssim_loss) + ld_opa * lv_opa + ld_scale * lscale
else:
loss = (1.0 - opt.lambda_dssim) * Ll1 + opt.lambda_dssim * (1.0 - ssim_loss) + ld_pdf * lv_pdf + ld_opa * lv_opa + ld_scale * lscale
loss += normal_loss + dist_loss + smooth_loss + tv_loss
if iteration % 2000 == 0 or iteration == 29999:
print("v_opa: {:.4f}".format(gaussians.Vth_opa.item()),
"v_pdf: {:.4f}".format(gaussians.get_Vth_pdf.mean().item()),
"num_gaussians:", gaussians._xyz.shape[0],
"loss: {:.4f}".format(loss.item()),
"norm: {:.4f}".format(normal_loss.item()),
"dist: {:.4f}".format(dist_loss.item()),
)
if iteration % 15000 == 0:
mask = gaussians.get_opacity < 0.5
print("mean_opac_low0.5:{:.4f}".format(gaussians.get_opacity[mask].mean().item()))
mask = gaussians.get_opacity < 0.9
print("mean_opac_low0.9:{:.4f}".format(gaussians.get_opacity[mask].mean().item()))
mask = gaussians.get_opacity < 1.01
print("mean_opac_low1.0:{:.4f}".format(gaussians.get_opacity[mask].mean().item()))
loss.backward()
iter_end.record()
with torch.no_grad():
# Progress bar
ema_loss_for_log = 0.4 * loss.item() + 0.6 * ema_loss_for_log
if iteration % 10 == 0:
progress_bar.set_postfix({"Loss": f"{ema_loss_for_log:.{7}f}"})
progress_bar.update(10)
if iteration == opt.iterations:
progress_bar.close()
# Log and save
training_report(tb_writer, iteration, Ll1, loss, l1_loss, iter_start.elapsed_time(iter_end), testing_iterations, scene, render, (pipe, background))
if (iteration in saving_iterations):
print("\n[ITER {}] Saving Gaussians".format(iteration))
scene.save(iteration)
# Densification
if iteration < opt.densify_until_iter:
# Keep track of max radii in image-space for pruning
gaussians.max_radii2D[visibility_filter] = torch.max(gaussians.max_radii2D[visibility_filter], radii[visibility_filter])
gaussians.add_densification_stats(viewspace_point_tensor, visibility_filter)
if iteration > opt.densify_from_iter and iteration % opt.densification_interval == 0:
size_threshold = 20 if iteration > opt.opacity_reset_interval else None
gaussians.densify_and_prune(opt.densify_grad_threshold, gaussians.Vth_opa.item(), scene.cameras_extent, size_threshold, radii, visibility_filter)
if (iteration % opt.opacity_reset_interval == 0 or (dataset.white_background and iteration == opt.densify_from_iter)) :
if pipe.no_spike:
gaussians.reset_opacity()
else:
gaussians.reset_opacity_spike()
elif not pipe.no_spike:
if iteration % opt.densification_interval == 0:
gaussians.spike_prune(gaussians.Vth_opa.item())
if iteration == 29999:
gaussians.last_prune(gaussians.Vth_opa.item())
# scale-based clone
if iteration >= 15000 and iteration <= 25000 and iteration % 800 == 0 and not pipe.no_spike:
gaussians.densify_and_prune_scale(opt.densify_grad_threshold, gaussians.Vth_opa.item(), scene.cameras_extent, size_threshold, radii, visibility_filter, scale_reg)
# if iteration >= 25000 and iteration % 200 == 0 and not pipe.no_spike:
# size_threshold = 20 if iteration > opt.opacity_reset_interval else None
# gaussians.prune_big(scene.cameras_extent, size_threshold)
# Optimizer step
if iteration < opt.iterations:
gaussians.optimizer.step()
gaussians.optimizer.zero_grad(set_to_none = True)
if (iteration in checkpoint_iterations):
print("\n[ITER {}] Saving Checkpoint".format(iteration))
torch.save((gaussians.capture(), iteration), scene.model_path + "/chkpnt" + str(iteration) + ".pth")
def prepare_output_and_logger(args):
if not args.model_path:
if os.getenv('OAR_JOB_ID'):
unique_str=os.getenv('OAR_JOB_ID')
else:
unique_str = str(uuid.uuid4())
args.model_path = os.path.join("./output/", unique_str[0:10])
# Set up output folder
print("Output folder: {}".format(args.model_path))
os.makedirs(args.model_path, exist_ok = True)
with open(os.path.join(args.model_path, "cfg_args"), 'w') as cfg_log_f:
cfg_log_f.write(str(Namespace(**vars(args))))
# Create Tensorboard writer
tb_writer = None
if TENSORBOARD_FOUND:
tb_writer = SummaryWriter(args.model_path)
else:
print("Tensorboard not available: not logging progress")
return tb_writer
def training_report(tb_writer, iteration, Ll1, loss, l1_loss, elapsed, testing_iterations, scene : Scene, renderFunc, renderArgs):
if tb_writer:
tb_writer.add_scalar('train_loss_patches/l1_loss', Ll1.item(), iteration)
tb_writer.add_scalar('train_loss_patches/total_loss', loss.item(), iteration)
tb_writer.add_scalar('iter_time', elapsed, iteration)
# Report test and samples of training set
if iteration in testing_iterations:
torch.cuda.empty_cache()
validation_configs = ({'name': 'test', 'cameras' : scene.getTestCameras()},
{'name': 'train', 'cameras' : [scene.getTrainCameras()[idx % len(scene.getTrainCameras())] for idx in range(5, 30, 5)]})
for config in validation_configs:
if config['cameras'] and len(config['cameras']) > 0:
l1_test = 0.0
psnr_test = 0.0
for idx, viewpoint in enumerate(config['cameras']):
image = torch.clamp(renderFunc(viewpoint, scene.gaussians, *renderArgs)["render"], 0.0, 1.0)
gt_image = torch.clamp(viewpoint.original_image.to("cuda"), 0.0, 1.0)
if tb_writer and (idx < 5):
tb_writer.add_images(config['name'] + "_view_{}/render".format(viewpoint.image_name), image[None], global_step=iteration)
if iteration == testing_iterations[0]:
tb_writer.add_images(config['name'] + "_view_{}/ground_truth".format(viewpoint.image_name), gt_image[None], global_step=iteration)
l1_test += l1_loss(image, gt_image).mean().double()
psnr_test += psnr(image, gt_image).mean().double()
psnr_test /= len(config['cameras'])
l1_test /= len(config['cameras'])
print("\n[ITER {}] Evaluating {}: L1 {} PSNR {}".format(iteration, config['name'], l1_test, psnr_test))
if tb_writer:
tb_writer.add_scalar(config['name'] + '/loss_viewpoint - l1_loss', l1_test, iteration)
tb_writer.add_scalar(config['name'] + '/loss_viewpoint - psnr', psnr_test, iteration)
if tb_writer:
tb_writer.add_histogram("scene/opacity_histogram", scene.gaussians.get_opacity, iteration)
tb_writer.add_scalar('total_points', scene.gaussians.get_xyz.shape[0], iteration)
torch.cuda.empty_cache()
if __name__ == "__main__":
# Set up command line argument parser
parser = ArgumentParser(description="Training script parameters")
lp = ModelParams(parser)
op = OptimizationParams(parser)
pp = PipelineParams(parser)
parser.add_argument('--ip', type=str, default="127.0.0.1")
parser.add_argument('--port', type=int, default=6009)
parser.add_argument('--debug_from', type=int, default=-1)
parser.add_argument('--detect_anomaly', action='store_true', default=False)
parser.add_argument("--test_iterations", nargs="+", type=int, default=[5_000, 10000, 20000, 30_000])
parser.add_argument("--save_iterations", nargs="+", type=int, default=[5_000, 10000, 20000, 30_000])
parser.add_argument("--quiet", action="store_true")
parser.add_argument("--checkpoint_iterations", nargs="+", type=int, default=[30000])
parser.add_argument("--start_checkpoint", type=str, default = None)
parser.add_argument('--scale', type=float, default=0.01)
parser.add_argument('--lambda_opa', type=float, default=0.00002)
parser.add_argument('--mask_normal', type=int, default=0)
args = parser.parse_args(sys.argv[1:])
args.save_iterations.append(args.iterations)
print("Optimizing " + args.model_path)
torch.manual_seed(100)
torch.cuda.manual_seed(100)
torch.backends.cudnn.deterministic = True
torch.backends.cudnn.benchmark = False
seed(100)
psnr = PeakSignalNoiseRatio().cuda()
# Initialize system state (RNG)
safe_state(args.quiet)
print(pp.extract(args).exp_name)
# Start GUI server, configure and run training
# network_gui.init(args.ip, args.port)
torch.autograd.set_detect_anomaly(args.detect_anomaly)
training(lp.extract(args), op.extract(args), pp.extract(args), args.test_iterations, args.save_iterations, args.checkpoint_iterations, args.start_checkpoint, args.debug_from, args.scale, args.lambda_opa, args.mask_normal)
# All done
print("\nTraining complete.")