-
Notifications
You must be signed in to change notification settings - Fork 182
/
bakedsdf.py
312 lines (265 loc) · 12.9 KB
/
bakedsdf.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
# Copyright 2022 The Nerfstudio Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
Implementation of BakedSDF.
"""
from __future__ import annotations
from dataclasses import dataclass, field
from typing import Dict, List, Tuple, Type
import numpy as np
import torch
from torch.nn import Parameter
from nerfstudio.cameras.rays import RayBundle
from nerfstudio.engine.callbacks import (
TrainingCallback,
TrainingCallbackAttributes,
TrainingCallbackLocation,
)
from nerfstudio.field_components.field_heads import FieldHeadNames
from nerfstudio.fields.density_fields import HashMLPDensityField
from nerfstudio.model_components.losses import interlevel_loss
from nerfstudio.model_components.ray_samplers import ProposalNetworkSampler
from nerfstudio.models.volsdf import VolSDFModel, VolSDFModelConfig
from nerfstudio.utils import colormaps
@dataclass
class BakedSDFModelConfig(VolSDFModelConfig):
"""BakedSDF Model Config"""
_target: Type = field(default_factory=lambda: BakedSDFFactoModel)
num_proposal_samples_per_ray: Tuple[int, ...] = (256, 96)
"""Number of samples per ray for the proposal network."""
num_neus_samples_per_ray: int = 48
"""Number of samples per ray for the nerf network."""
proposal_update_every: int = 5
"""Sample every n steps after the warmup"""
proposal_warmup: int = 5000
"""Scales n from 1 to proposal_update_every over this many steps"""
num_proposal_iterations: int = 2
"""Number of proposal network iterations."""
use_same_proposal_network: bool = False
"""Use the same proposal network. Otherwise use different ones."""
proposal_net_args_list: List[Dict] = field(
default_factory=lambda: [
{"hidden_dim": 16, "log2_hashmap_size": 17, "num_levels": 5, "max_res": 64},
{"hidden_dim": 16, "log2_hashmap_size": 17, "num_levels": 5, "max_res": 256},
]
)
"""Arguments for the proposal density fields."""
interlevel_loss_mult: float = 1.0
"""Proposal loss multiplier."""
use_proposal_weight_anneal: bool = True
"""Whether to use proposal weight annealing."""
proposal_weights_anneal_slope: float = 10.0
"""Slope of the annealing function for the proposal weights."""
proposal_weights_anneal_max_num_iters: int = 1000
"""Max num iterations for the annealing function."""
use_single_jitter: bool = True
"""Whether use single jitter or not for the proposal networks."""
use_anneal_beta: bool = True
"""whether to use anneal beta"""
beta_anneal_max_num_iters: int = 250000
"""Max num iterations for the annealing of beta in laplacian density."""
beta_anneal_init: float = 0.1
"""Initial value of beta for the annealing of beta in laplacian density."""
beta_anneal_end: float = 0.001
"""End value of beta for the annealing of beta in laplacian density."""
use_anneal_eikonal_weight: bool = False
"""whether to use annealing for eikonal loss weight"""
eikonal_anneal_max_num_iters: int = 250000
"""Max num iterations for the annealing of beta in laplacian density."""
use_spatial_varying_eikonal_loss: bool = False
"""whether to use different weight of eikonal loss based the points norm, farway points have large weights"""
eikonal_loss_mult_start: float = 0.01
eikonal_loss_mult_end: float = 0.1
eikonal_loss_mult_slop: float = 2.0
class BakedSDFFactoModel(VolSDFModel):
"""BakedSDF model
Args:
config: BakedSDF configuration to instantiate model
"""
config: BakedSDFModelConfig
def populate_modules(self):
"""Set the fields and modules."""
super().populate_modules()
self.density_fns = []
num_prop_nets = self.config.num_proposal_iterations
# Build the proposal network(s)
self.proposal_networks = torch.nn.ModuleList()
if self.config.use_same_proposal_network:
assert len(self.config.proposal_net_args_list) == 1, "Only one proposal network is allowed."
prop_net_args = self.config.proposal_net_args_list[0]
network = HashMLPDensityField(
self.scene_box.aabb, spatial_distortion=self.scene_contraction, **prop_net_args
)
self.proposal_networks.append(network)
self.density_fns.extend([network.density_fn for _ in range(num_prop_nets)])
else:
for i in range(num_prop_nets):
prop_net_args = self.config.proposal_net_args_list[min(i, len(self.config.proposal_net_args_list) - 1)]
network = HashMLPDensityField(
self.scene_box.aabb,
spatial_distortion=self.scene_contraction,
**prop_net_args,
)
self.proposal_networks.append(network)
self.density_fns.extend([network.density_fn for network in self.proposal_networks])
# update proposal network every iterations
update_schedule = lambda step: -1
self.proposal_sampler = ProposalNetworkSampler(
num_nerf_samples_per_ray=self.config.num_neus_samples_per_ray,
num_proposal_samples_per_ray=self.config.num_proposal_samples_per_ray,
num_proposal_network_iterations=self.config.num_proposal_iterations,
use_uniform_sampler=False,
single_jitter=self.config.use_single_jitter,
update_sched=update_schedule,
)
def get_param_groups(self) -> Dict[str, List[Parameter]]:
param_groups = {}
if self.config.use_anneal_beta:
# don't optimize beta in laplace density if use annealing beta
param_groups["fields"] = [
n_p[1] for n_p in filter(lambda n_p: "laplace_density" not in n_p[0], self.field.named_parameters())
]
else:
param_groups["fields"] = list(self.field.parameters())
param_groups["proposal_networks"] = list(self.proposal_networks.parameters())
if self.config.background_model != "none":
param_groups["field_background"] = list(self.field_background.parameters())
else:
param_groups["field_background"] = list(self.field_background)
return param_groups
def get_training_callbacks(
self, training_callback_attributes: TrainingCallbackAttributes
) -> List[TrainingCallback]:
callbacks = super().get_training_callbacks(training_callback_attributes)
if self.config.use_proposal_weight_anneal:
# anneal the weights of the proposal network before doing PDF sampling
N = self.config.proposal_weights_anneal_max_num_iters
def set_anneal(step):
# https://arxiv.org/pdf/2111.12077.pdf eq. 18
train_frac = np.clip(step / N, 0, 1)
bias = lambda x, b: (b * x) / ((b - 1) * x + 1)
anneal = bias(train_frac, self.config.proposal_weights_anneal_slope)
self.proposal_sampler.set_anneal(anneal)
callbacks.append(
TrainingCallback(
where_to_run=[TrainingCallbackLocation.BEFORE_TRAIN_ITERATION],
update_every_num_iters=1,
func=set_anneal,
)
)
callbacks.append(
TrainingCallback(
where_to_run=[TrainingCallbackLocation.AFTER_TRAIN_ITERATION],
update_every_num_iters=1,
func=self.proposal_sampler.step_cb,
)
)
if self.config.use_anneal_beta:
# anneal the beta of volsdf before each training iterations
M = self.config.beta_anneal_max_num_iters
beta_init = self.config.beta_anneal_init
beta_end = self.config.beta_anneal_end
def set_beta(step):
# bakedsdf's beta schedule
train_frac = np.clip(step / M, 0, 1)
beta = beta_init / (1 + (beta_init - beta_end) / beta_end * (train_frac**0.8))
self.field.laplace_density.beta.data[...] = beta
callbacks.append(
TrainingCallback(
where_to_run=[TrainingCallbackLocation.BEFORE_TRAIN_ITERATION],
update_every_num_iters=1,
func=set_beta,
)
)
if self.config.use_anneal_eikonal_weight:
# anneal the beta of volsdf before each training iterations
K = self.config.eikonal_anneal_max_num_iters
weight_init = 0.01
weight_end = 0.1
def set_weight(step):
# bakedsdf's beta schedule
train_frac = np.clip(step / K, 0, 1)
mult = weight_end / (1 + (weight_end - weight_init) / weight_init * ((1.0 - train_frac) ** 10))
self.config.eikonal_loss_mult = mult
callbacks.append(
TrainingCallback(
where_to_run=[TrainingCallbackLocation.BEFORE_TRAIN_ITERATION],
update_every_num_iters=1,
func=set_weight,
)
)
return callbacks
def sample_and_forward_field(self, ray_bundle: RayBundle):
ray_samples, weights_list, ray_samples_list = self.proposal_sampler(ray_bundle, density_fns=self.density_fns)
# TODO only forward the points that are inside the sphere
field_outputs = self.field(ray_samples)
field_outputs[FieldHeadNames.ALPHA] = ray_samples.get_alphas(field_outputs[FieldHeadNames.DENSITY])
if self.config.background_model != "none":
field_outputs = self.forward_background_field_and_merge(ray_samples, field_outputs)
weights = ray_samples.get_weights_from_alphas(field_outputs[FieldHeadNames.ALPHA])
weights_list.append(weights)
ray_samples_list.append(ray_samples)
samples_and_field_outputs = {
"ray_samples": ray_samples,
"field_outputs": field_outputs,
"weights": weights,
"weights_list": weights_list,
"ray_samples_list": ray_samples_list,
}
return samples_and_field_outputs
def get_loss_dict(self, outputs, batch, metrics_dict=None):
loss_dict = {}
image = batch["image"].to(self.device)
loss_dict["rgb_loss"] = self.rgb_loss(image, outputs["rgb"])
if self.training:
# eikonal loss
grad_theta = outputs["eik_grad"]
# s3im loss
if self.config.s3im_loss_mult > 0:
loss_dict["s3im_loss"] = self.s3im_loss(image, outputs["rgb"]) * self.config.s3im_loss_mult
if self.config.use_spatial_varying_eikonal_loss:
points_norm = outputs["points_norm"][..., 0]
points_weights = torch.where(points_norm <= 1, torch.ones_like(points_norm), points_norm)
# shortcut
weight_init = self.config.eikonal_loss_mult_start
weight_end = self.config.eikonal_loss_mult_end
slop = self.config.eikonal_loss_mult_slop
points_weights = weight_end / (
1 + (weight_end - weight_init) / weight_init * ((2.0 - points_weights) ** slop)
)
loss_dict["eikonal_loss"] = (((grad_theta.norm(2, dim=-1) - 1) ** 2) * points_weights).mean()
else:
loss_dict["eikonal_loss"] = (
(grad_theta.norm(2, dim=-1) - 1) ** 2
).mean() * self.config.eikonal_loss_mult
loss_dict["interlevel_loss"] = self.config.interlevel_loss_mult * interlevel_loss(
outputs["weights_list"], outputs["ray_samples_list"]
)
return loss_dict
def get_image_metrics_and_images(
self, outputs: Dict[str, torch.Tensor], batch: Dict[str, torch.Tensor]
) -> Tuple[Dict[str, float], Dict[str, torch.Tensor]]:
metrics_dict, images_dict = super().get_image_metrics_and_images(outputs, batch)
for i in range(self.config.num_proposal_iterations):
key = f"prop_depth_{i}"
prop_depth_i = colormaps.apply_depth_colormap(
outputs[key],
accumulation=outputs["accumulation"],
)
images_dict[key] = prop_depth_i
return metrics_dict, images_dict
def get_metrics_dict(self, outputs, batch) -> Dict:
metric_dict = super().get_metrics_dict(outputs, batch)
metric_dict["eikonal_loss_mult"] = self.config.eikonal_loss_mult
return metric_dict