File size: 12,506 Bytes
7ab05dd | 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 | import torch
import torch.nn as nn
import torch_scatter
import torch_cluster
from peft import LoraConfig, get_peft_model
from collections import OrderedDict
from pointcept.models.losses import build_criteria
from pointcept.models.utils.structure import Point
from pointcept.models.utils import offset2batch
from .builder import MODELS, build_model
@MODELS.register_module()
class DefaultSegmentor(nn.Module):
def __init__(self, backbone=None, criteria=None):
super().__init__()
self.backbone = build_model(backbone)
self.criteria = build_criteria(criteria)
def forward(self, input_dict):
if "condition" in input_dict.keys():
# PPT (https://arxiv.org/abs/2308.09718)
# currently, only support one batch one condition
input_dict["condition"] = input_dict["condition"][0]
seg_logits = self.backbone(input_dict)
# train
if self.training:
loss = self.criteria(seg_logits, input_dict["segment"])
return dict(loss=loss)
# eval
elif "segment" in input_dict.keys():
loss = self.criteria(seg_logits, input_dict["segment"])
return dict(loss=loss, seg_logits=seg_logits)
# test
else:
return dict(seg_logits=seg_logits)
@MODELS.register_module()
class DefaultSegmentorV2(nn.Module):
def __init__(
self,
num_classes,
backbone_out_channels,
backbone=None,
criteria=None,
freeze_backbone=False,
):
super().__init__()
self.seg_head = (
nn.Linear(backbone_out_channels, num_classes)
if num_classes > 0
else nn.Identity()
)
self.backbone = build_model(backbone)
self.criteria = build_criteria(criteria)
self.freeze_backbone = freeze_backbone
if self.freeze_backbone:
for p in self.backbone.parameters():
p.requires_grad = False
def forward(self, input_dict, return_point=False):
point = Point(input_dict)
point = self.backbone(point)
# Backbone added after v1.5.0 return Point instead of feat and use DefaultSegmentorV2
# TODO: remove this part after make all backbone return Point only.
if isinstance(point, Point):
while "pooling_parent" in point.keys():
assert "pooling_inverse" in point.keys()
parent = point.pop("pooling_parent")
inverse = point.pop("pooling_inverse")
parent.feat = torch.cat([parent.feat, point.feat[inverse]], dim=-1)
point = parent
feat = point.feat
else:
feat = point
seg_logits = self.seg_head(feat)
return_dict = dict()
if return_point:
# PCA evaluator parse feat and coord in point
return_dict["point"] = point
# train
if self.training:
loss = self.criteria(seg_logits, input_dict["segment"])
return_dict["loss"] = loss
# eval
elif "segment" in input_dict.keys():
loss = self.criteria(seg_logits, input_dict["segment"])
return_dict["loss"] = loss
return_dict["seg_logits"] = seg_logits
# test
else:
return_dict["seg_logits"] = seg_logits
return return_dict
@MODELS.register_module()
class DefaultLORASegmentorV2(nn.Module):
def __init__(
self,
num_classes,
backbone_out_channels,
backbone=None,
criteria=None,
freeze_backbone=False,
use_lora=False,
lora_r=8,
lora_alpha=16,
lora_dropout=0.1,
backbone_path=None,
keywords=None,
replacements=None,
):
super().__init__()
self.seg_head = (
nn.Linear(backbone_out_channels, num_classes)
if num_classes > 0
else nn.Identity()
)
self.keywords = keywords
self.replacements = replacements
self.backbone = build_model(backbone)
backbone_weight = torch.load(
backbone_path,
map_location=lambda storage, loc: storage.cuda(),
)
self.backbone_load(backbone_weight)
self.criteria = build_criteria(criteria)
self.freeze_backbone = freeze_backbone
self.use_lora = use_lora
if self.use_lora:
lora_config = LoraConfig(
r=lora_r,
lora_alpha=lora_alpha,
target_modules=["qkv"],
# target_modules=["query", "value"],
lora_dropout=lora_dropout,
bias="none",
)
self.backbone.enc = get_peft_model(self.backbone.enc, lora_config)
if self.freeze_backbone:
for p in self.backbone.parameters():
p.requires_grad = False
if self.use_lora:
for name, param in self.backbone.named_parameters():
if "lora_" in name:
param.requires_grad = True
self.backbone.enc.print_trainable_parameters()
def backbone_load(self, checkpoint):
weight = OrderedDict()
for key, value in checkpoint["state_dict"].items():
if not key.startswith("module."):
key = "module." + key # xxx.xxx -> module.xxx.xxx
# Now all keys contain "module." no matter DDP or not.
if self.keywords in key:
key = key.replace(self.keywords, self.replacements)
key = key[7:] # module.xxx.xxx -> xxx.xxx
if key.startswith("backbone."):
key = key[9:]
weight[key] = value
load_state_info = self.backbone.load_state_dict(weight, strict=False)
print(f"Missing keys: {load_state_info[0]}")
print(f"Unexpected keys: {load_state_info[1]}")
def forward(self, input_dict, return_point=False):
point = Point(input_dict)
if self.freeze_backbone and not self.use_lora:
with torch.no_grad():
point = self.backbone(point)
else:
point = self.backbone(point)
if isinstance(point, Point):
while "pooling_parent" in point.keys():
assert "pooling_inverse" in point.keys()
parent = point.pop("pooling_parent")
inverse = point.pop("pooling_inverse")
parent.feat = torch.cat([parent.feat, point.feat[inverse]], dim=-1)
point = parent
feat = point.feat
else:
feat = point
seg_logits = self.seg_head(feat)
return_dict = dict()
if return_point:
return_dict["point"] = point
if self.training:
loss = self.criteria(seg_logits, input_dict["segment"])
return_dict["loss"] = loss
elif "segment" in input_dict.keys():
loss = self.criteria(seg_logits, input_dict["segment"])
return_dict["loss"] = loss
return_dict["seg_logits"] = seg_logits
else:
return_dict["seg_logits"] = seg_logits
return return_dict
@MODELS.register_module()
class DINOEnhancedSegmentor(nn.Module):
def __init__(
self,
num_classes,
backbone_out_channels,
backbone=None,
criteria=None,
freeze_backbone=False,
):
super().__init__()
self.seg_head = (
nn.Linear(backbone_out_channels, num_classes)
if num_classes > 0
else nn.Identity()
)
self.backbone = build_model(backbone) if backbone is not None else None
self.criteria = build_criteria(criteria)
self.freeze_backbone = freeze_backbone
if self.backbone is not None and self.freeze_backbone:
for p in self.backbone.parameters():
p.requires_grad = False
def forward(self, input_dict, return_point=False):
point = Point(input_dict)
if self.backbone is not None:
if self.freeze_backbone:
with torch.no_grad():
point = self.backbone(point)
else:
point = self.backbone(point)
point_list = [point]
while "unpooling_parent" in point_list[-1].keys():
point_list.append(point_list[-1].pop("unpooling_parent"))
for i in reversed(range(1, len(point_list))):
point = point_list[i]
parent = point_list[i - 1]
assert "pooling_inverse" in point.keys()
inverse = point.pooling_inverse
parent.feat = torch.cat([parent.feat, point.feat[inverse]], dim=-1)
point = point_list[0]
while "pooling_parent" in point.keys():
assert "pooling_inverse" in point.keys()
parent = point.pop("pooling_parent")
inverse = point.pooling_inverse
parent.feat = torch.cat([parent.feat, point.feat[inverse]], dim=-1)
point = parent
feat = [point.feat]
else:
feat = []
dino_coord = input_dict["dino_coord"]
dino_feat = input_dict["dino_feat"]
dino_offset = input_dict["dino_offset"]
idx = torch_cluster.knn(
x=dino_coord,
y=point.origin_coord,
batch_x=offset2batch(dino_offset),
batch_y=offset2batch(point.origin_offset),
k=1,
)[1]
feat.append(dino_feat[idx])
feat = torch.concatenate(feat, dim=-1)
seg_logits = self.seg_head(feat)
return_dict = dict()
if return_point:
# PCA evaluator parse feat and coord in point
return_dict["point"] = point
# train
if self.training:
loss = self.criteria(seg_logits, input_dict["origin_segment"])
return_dict["loss"] = loss
# eval
elif "segment" in input_dict.keys():
loss = self.criteria(seg_logits, input_dict["origin_segment"])
return_dict["loss"] = loss
return_dict["seg_logits"] = seg_logits
# test
else:
return_dict["seg_logits"] = seg_logits
return return_dict
@MODELS.register_module()
class DefaultClassifier(nn.Module):
def __init__(
self,
backbone=None,
criteria=None,
num_classes=40,
backbone_embed_dim=256,
freeze_backbone=False,
):
super().__init__()
self.backbone = build_model(backbone)
self.criteria = build_criteria(criteria)
self.num_classes = num_classes
self.backbone_embed_dim = backbone_embed_dim
self.freeze_backbone = freeze_backbone
self.cls_head = nn.Sequential(
nn.Linear(backbone_embed_dim, 256),
nn.BatchNorm1d(256),
nn.ReLU(inplace=True),
nn.Dropout(p=0.5),
nn.Linear(256, 128),
nn.BatchNorm1d(128),
nn.ReLU(inplace=True),
nn.Dropout(p=0.5),
nn.Linear(128, num_classes),
)
if self.freeze_backbone:
for p in self.backbone.parameters():
p.requires_grad = False
self.backbone.eval()
def forward(self, input_dict):
point = Point(input_dict)
if self.freeze_backbone:
with torch.no_grad():
point = self.backbone(point)
else:
point = self.backbone(point)
# Backbone added after v1.5.0 return Point instead of feat
# And after v1.5.0 feature aggregation for classification operated in classifier
# TODO: remove this part after make all backbone return Point only.
if isinstance(point, Point):
point.feat = torch_scatter.segment_csr(
src=point.feat,
indptr=nn.functional.pad(point.offset, (1, 0)),
reduce="mean",
)
feat = point.feat
else:
feat = point
cls_logits = self.cls_head(feat)
if self.training:
loss = self.criteria(cls_logits, input_dict["category"])
return dict(loss=loss)
elif "category" in input_dict.keys():
loss = self.criteria(cls_logits, input_dict["category"])
return dict(loss=loss, cls_logits=cls_logits)
else:
return dict(cls_logits=cls_logits)
|