Repository navigation
Expand file tree
/
Copy pathtrain_stage1.py
More file actions
executable file
·393 lines (343 loc) · 19.1 KB
/
Copy pathtrain_stage1.py
File metadata and controls
executable file
·393 lines (343 loc) · 19.1 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
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
"""
stage1 : vanilla 3DGS training
"""
#
# 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 george.drettakis@inria.fr
#
import os
import torch
from random import randint
from utils.loss_utils import l1_loss, ssim, get_expon_lr_func
from gaussian_renderer import network_gui, render_original, render_upgrade
import sys
from scene import Scene, GaussianModel
from utils.general_utils import safe_state
import uuid
from tqdm import tqdm
from utils.image_utils import psnr
from argparse import ArgumentParser, Namespace
from arguments import ModelParams, PipelineParams, OptimizationParams, MaskingParams
try:
from torch.utils.tensorboard import SummaryWriter
TENSORBOARD_FOUND = True
except ImportError:
TENSORBOARD_FOUND = False
import wandb
from omegaconf import OmegaConf
from omegaconf.dictconfig import DictConfig
from torchvision.transforms import ToPILImage
import torchvision
import numpy as np
def set_render_func(opt):
app_model = opt.appearance_model
use_invdepth = opt.inv_depth # always True
if app_model == "expmaps" or use_invdepth:
print("[INFO] : using upgraded-3DGS render")
return render_upgrade
elif app_model == "original":
print("[INFO] : using original render")
return render_original
elif app_model == "hier":
raise NotImplementedError("hier render not implemented")
else:
print("[INFO] : using default render")
return render
def training(dataset, opt, pipe, mask_params, testing_iterations, saving_iterations, checkpoint_iterations, checkpoint, debug_from, logger=None, args=None):
# Settings -------------------------------------------------------------- #
first_iter = 0
tb_writer = prepare_output_and_logger(dataset, opt, mask_params)
# else:
# tb_writer = None
tb_writer = None
print(f"[INFO] : args.swap = {args.swap}")
if args.swap:
dataset.dataset_distribution = "swap"
gaussians = GaussianModel(dataset, opt, mask_params)
# gaussians = GaussianModelHybrid(dataset, opt, mask_params) if opt.appearance_model == "featsplat" else GaussianModel(dataset, opt, mask_params)
scene = Scene(dataset, mask_params, gaussians)
render_func = set_render_func(opt)
print(render_func)
early_stop = 10000 if args.earlystop else False
image_log_interval = args.image_log_interval
if early_stop:
print(f"[INFO] : training early stops at {early_stop}")
gaussians.training_setup(opt, mask_params)
if checkpoint:
(model_params, first_iter) = torch.load(checkpoint)
gaussians.restore(model_params, opt, mask_params)
bg_color = [1, 1, 1] if dataset.white_background else [0, 0, 0]
background = torch.tensor(bg_color, dtype=torch.float32, device="cuda")
# Depth regularization ; from updated 3DGS
depth_l1_weight = get_expon_lr_func(opt.depth_l1_weight_init, opt.depth_l1_weight_final, max_steps=opt.iterations)
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
ema_L1depth_for_log = 0.0
progress_bar = tqdm(range(first_iter, opt.iterations), desc="Training progress")
first_iter += 1
for iteration in range(first_iter, opt.iterations + 1):
# For debugging
# import time
# time.sleep(iteration*2)
# if iteration > 50:
# break
if early_stop: # quite hardcoded
if iteration > 1000:
break
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_func(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)
gaussians.update_sh_degree(iteration)
# # Every 1000 its we increase the levels of SH up to a maximum degree
if iteration % 1000 == 0: # TODO: slowing down maybe helpful
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))
# viewpoint_cam = viewpoint_stack[0]
# Render
if (iteration - 1) == debug_from:
pipe.debug = True
bg = torch.rand((3), device="cuda") if opt.random_background else background
render_pkg = render_func(viewpoint_cam, gaussians, pipe, bg)
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 = l1_loss(image, gt_image)
loss = (1.0 - opt.lambda_dssim) * Ll1 + opt.lambda_dssim * (1.0 - ssim(image, gt_image))
# Depth regularization
Ll1depth_pure = 0.0
if depth_l1_weight(iteration) > 0 and viewpoint_cam.depth_reliable:
invDepth = render_pkg["invdepth"][0, :, :][None] # accumulated depth
mono_invdepth = viewpoint_cam.invdepthmap.cuda()
depth_mask = viewpoint_cam.depth_mask.cuda()
# exit()
Ll1depth_pure = torch.abs((invDepth - mono_invdepth) * depth_mask).mean()
Ll1depth = depth_l1_weight(iteration) * Ll1depth_pure
loss += Ll1depth
Ll1depth = Ll1depth.item()
else:
Ll1depth = 0
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
ema_L1depth_for_log = 0.4 * Ll1depth + 0.6 * ema_L1depth_for_log
if iteration % 10 == 0:
progress_bar.set_postfix({"Loss": f"{ema_loss_for_log:.{7}f}", "Depth Loss": f"{ema_L1depth_for_log:.{7}f}"})
progress_bar.update(10)
if iteration == opt.iterations:
progress_bar.close()
# Log and save
if tb_writer is not None:
training_report(tb_writer, iteration, Ll1, loss, l1_loss, iter_start.elapsed_time(iter_end), testing_iterations, scene, render_func, (pipe, background))
else:
training_report_wandb(logger, iteration, Ll1, loss, l1_loss, iter_start.elapsed_time(iter_end), testing_iterations, scene, render_func, (pipe, background),
image_log_interval, scene.model_path, dataset, mask_params)
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, opt.min_opacity, scene.cameras_extent, size_threshold)
if iteration % opt.opacity_reset_interval == 0 or (dataset.white_background and iteration == opt.densify_from_iter):
gaussians.reset_opacity()
# Optimizer step
if iteration < opt.iterations:
if opt.appearance_model == "expmaps":
gaussians.exposure_optimizer.step()
gaussians.exposure_optimizer.zero_grad(set_to_none = True)
gaussians.optimizer.step()
gaussians.optimizer.zero_grad(set_to_none = True)
# print(f"[DEBUG] : GPU MEM CHECK : {iteration:04d}th iteration after optim.step = {torch.cuda.memory_allocated() / 1024**2:.2f} MB")
# gaussians.vmf_mu_normalize() # only act when use_vmf = 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, opt, mask_params):
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))
merged_args = {**vars(args), **vars(opt), **vars(mask_params)}
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))))
cfg_log_f.write(str(Namespace(**merged_args))) # MaskParams should be also saved
# 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()
def training_report_wandb(logger, iteration, Ll1, loss, l1_loss, elapsed, testing_iterations, scene : Scene, renderFunc, renderArgs, image_log_interval, model_path, dataset, mask_params, additional_loss=None):
report_dict = {
"train_loss_patches/l1_loss":Ll1.item(),
"train_loss_patches/total_loss": loss.item(),
# "train_loss_patches/sds_loss": sds_loss.item(),
"iter_time":elapsed
}
logger.log(report_dict, step=iteration)
if iteration % image_log_interval == 0:
render_path = os.path.join(model_path, f"log-{iteration:04d}")
os.makedirs(render_path, exist_ok=True)
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']):
# render_dict = renderFunc(viewpoint, scene.gaussians, *renderArgs)
# render_dict = renderFunc(viewpoint, scene.gaussians, pipe=renderArgs[0], bg_color=renderArgs[1], vmf_evaluate=False)
render_dict = renderFunc(viewpoint, scene.gaussians, pipe=renderArgs[0], bg_color=renderArgs[1])
image = torch.clamp(render_dict["render"], 0.0, 1.0)
gt_image = torch.clamp(viewpoint.original_image.to("cuda"), 0.0, 1.0)
use_depth = viewpoint.invdepthmap is not None
if use_depth:
invdepth = torch.clamp(render_dict["invdepth"][0, :, :][None], 0.0, 1.0)
gt_invdepth = torch.clamp(viewpoint.invdepthmap.cuda(), 0.0, 1.0)
prefix = config["name"] + "-render-" + viewpoint.image_name
if iteration >= 100:
torchvision.utils.save_image(image, os.path.join(render_path, prefix + ".png"))
if use_depth:
torchvision.utils.save_image(invdepth, os.path.join(render_path, prefix + "-invdepth.png"))
# save_mask_rendered_images(mask_params, render_dict, render_path, prefix, viewpoint)
# skipping gt
if iteration == image_log_interval:
torchvision.utils.save_image(gt_image, os.path.join(render_path, "gt-" + viewpoint.image_name + ".png"))
if use_depth:
torchvision.utils.save_image(gt_invdepth, os.path.join(render_path, "gt-invdepth-" + viewpoint.image_name + ".png"))
if iteration in testing_iterations:
l1_test += l1_loss(image, gt_image).mean().double()
psnr_test += psnr(image, gt_image).mean().double()
else:
if idx > 10: break # prevent logging getting too heavy memory
if iteration in testing_iterations:
psnr_test /= len(config['cameras'])
l1_test /= len(config['cameras'])
print("\n[ITER {}] Evaluating {}: L1 {} PSNR {}".format(iteration, config['name'], l1_test, psnr_test))
wandb.log({f"{config['name']}/loss_viewpoint_l1-loss":l1_test}, step=iteration)
wandb.log({f"{config['name']}/loss_viewpoint_psnr":psnr_test}, step=iteration)
wandb.log({"train/scene_opacity": wandb.Histogram(np_histogram=np.histogram(scene.gaussians.get_opacity.cpu().numpy()))}, step=iteration)
wandb.log({"train/total_points": scene.gaussians.get_xyz.shape[0]}, step=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)
mp = MaskingParams(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=[3_000, 7_000, 10_000, 15_000, 20_000, 30_000])
parser.add_argument("--save_iterations", nargs="+", type=int, default=[30_000])
parser.add_argument("--quiet", action="store_true")
parser.add_argument("--checkpoint_iterations", nargs="+", type=int, default=[30_000])
parser.add_argument("--start_checkpoint", type=str, default = None)
parser.add_argument("--config", "-c", type=str, default="")
parser.add_argument("--expname", type=str, default="")
parser.add_argument("--image_log_interval", type=int, default=5000)
parser.add_argument("--swap", action="store_true")
parser.add_argument("--earlystop", action="store_true")
# parser.add_argument("--no_sds", action="store_true")
args = parser.parse_args(sys.argv[1:])
args.save_iterations.append(args.iterations)
scene_name = os.path.basename(args.source_path)
def recursive_merge(key, host):
if isinstance(host[key], DictConfig):
for key1 in host[key].keys():
recursive_merge(key1, host[key])
else:
assert hasattr(args, key), key
setattr(args, key, host[key])
save_expname = "stage1_"
save_expname += args.config.split("/")[-1]
cfg = OmegaConf.load(args.config + ".yaml")
for k in cfg.keys():
recursive_merge(k, cfg)
args.model_path = os.path.join("output", save_expname, scene_name)
if args.swap:
args.model_path = os.path.join("output", save_expname, scene_name + "_swap")
print("Optimizing " + args.model_path)
logger = wandb.init(name=f"stage1_{args.expname}_{scene_name}", project='3dgs-diff', mode='offline')
# Initialize system state (RNG)
safe_state(args.quiet)
# 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), mp.extract(args), args.test_iterations, args.save_iterations, args.checkpoint_iterations, args.start_checkpoint, args.debug_from,
logger=logger, args=args)
# All done
print("\nTraining complete.")
wandb.finish()