I quote ghm loss in ssd, loss_bbox can normally drop to about 0.8, but loss_cls can only drop to about 3, what’s going on?code show as below：
import numpy as np
import torch
import torch.nn as nn
import torch.nn.functional as F
from mmcv.cnn import xavier_init
from mmdet.core import AnchorGenerator, anchor_target, multi_apply
from .anchor_head import AnchorHead
from …losses import smooth_l1_loss
from …registry import HEADS
def _expand_binary_labels(labels, label_weights, label_channels):
bin_labels = labels.new_full((labels.size(0), label_channels), 0)
inds = torch.nonzero(labels >= 1).squeeze()
if inds.numel() > 0:
bin_labels[inds, labels[inds] - 1] = 1
bin_label_weights = label_weights.view(-1, 1).expand(label_weights.size(0), label_channels)
return bin_labels, bin_label_weights
def GHM_Class_loss(pred, target, label_weight):
bins = 30
mmt = 0.75
loss_weight = 1.0
edges = torch.arange(bins + 1).float().cuda() / bins
edges[-1] += 1e-6
if mmt > 0:
acc_sum = torch.zeros(bins).cuda()
use_sigmoid = True
if not use_sigmoid:
raise NotImplementedError
if pred.dim() != target.dim():
target, label_weight = _expand_binary_labels(
target, label_weight, pred.size(-1))
target, label_weight = target.float(), label_weight.float()
weights = torch.zeros_like(pred)
# gradient length
g = torch.abs(pred.sigmoid().detach() - target)
valid = label_weight > 0
tot = max(valid.float().sum().item(), 1.0)
n = 0 # n valid bins
for i in range(bins):
inds = (g >= edges[i]) & (g < edges[i + 1]) & valid
num_in_bin = inds.sum().item()
if num_in_bin > 0:
if mmt > 0:
acc_sum[i] = mmt * acc_sum[i] + (1 - mmt) * num_in_bin
weights[inds] = tot / acc_sum[i]
else:
weights[inds] = tot / num_in_bin
n += 1
if n > 0:
weights = weights / n
loss = F.binary_cross_entropy_with_logits(
pred, target, weights, reduction='sum') / tot
return loss * loss_weight
def GHM_Reg_loss(pred, target, label_weight, avg_factor=None):
mu = 0.02
mmt = 0.70
bins = 10
loss_weight = 1.0
edges = torch.arange(bins + 1).float().cuda() / bins
edges[-1] = 1e3
if mmt > 0:
acc_sum = torch.zeros(bins).cuda()
# ASL1 loss
diff = pred - target
loss = torch.sqrt(diff * diff + mu * mu) - mu
# gradient length
g = torch.abs(diff / torch.sqrt(mu * mu + diff * diff)).detach()
weights = torch.zeros_like(g)
valid = label_weight > 0
tot = max(label_weight.float().sum().item(), 1.0)
n = 0 # n: valid bins
for i in range(bins):
inds = (g >= edges[i]) & (g < edges[i + 1]) & valid
num_in_bin = inds.sum().item()
if num_in_bin > 0:
n += 1
if mmt > 0:
acc_sum[i] = mmt * acc_sum[i] \
+ (1 - mmt) * num_in_bin
weights[inds] = tot / acc_sum[i]
else:
weights[inds] = tot / num_in_bin
if n > 0:
weights /= n
loss = loss * weights
loss = loss.sum() / tot
return loss * loss_weight
TODO: add loss evaluator for SSD
@HEADS.register_module
class SSDHead(AnchorHead):
def __init__(self,
input_size=300,
num_classes=81,
in_channels=(512, 1024, 512, 256, 256, 256),
anchor_strides=(8, 16, 32, 64, 100, 300),
basesize_ratio_range=(0.1, 0.9),
anchor_ratios=([2], [2, 3], [2, 3], [2, 3], [2], [2]),
target_means=(.0, .0, .0, .0),
target_stds=(1.0, 1.0, 1.0, 1.0)):
super(AnchorHead, self).__init__()
self.input_size = input_size
self.num_classes = num_classes
self.in_channels = in_channels
self.cls_out_channels = num_classes
num_anchors = [len(ratios) * 2 + 2 for ratios in anchor_ratios]
reg_convs = []
cls_convs = []
for i in range(len(in_channels)):
reg_convs.append(
nn.Conv2d(
in_channels[i],
num_anchors[i] * 4,
kernel_size=3,
padding=1))
cls_convs.append(
nn.Conv2d(
in_channels[i],
num_anchors[i] * num_classes,
kernel_size=3,
padding=1))
self.reg_convs = nn.ModuleList(reg_convs)
self.cls_convs = nn.ModuleList(cls_convs)
min_ratio, max_ratio = basesize_ratio_range
min_ratio = int(min_ratio * 100)
max_ratio = int(max_ratio * 100)
step = int(np.floor(max_ratio - min_ratio) / (len(in_channels) - 2))
min_sizes = []
max_sizes = []
for r in range(int(min_ratio), int(max_ratio) + 1, step):
min_sizes.append(int(input_size * r / 100))
max_sizes.append(int(input_size * (r + step) / 100))
if input_size == 300:
if basesize_ratio_range[0] == 0.15: # SSD300 COCO
min_sizes.insert(0, int(input_size * 7 / 100))
max_sizes.insert(0, int(input_size * 15 / 100))
elif basesize_ratio_range[0] == 0.2: # SSD300 VOC
min_sizes.insert(0, int(input_size * 10 / 100))
max_sizes.insert(0, int(input_size * 20 / 100))
elif input_size == 512:
if basesize_ratio_range[0] == 0.1: # SSD512 COCO
min_sizes.insert(0, int(input_size * 4 / 100))
max_sizes.insert(0, int(input_size * 10 / 100))
elif basesize_ratio_range[0] == 0.15: # SSD512 VOC
min_sizes.insert(0, int(input_size * 7 / 100))
max_sizes.insert(0, int(input_size * 15 / 100))
self.anchor_generators = []
self.anchor_strides = anchor_strides
for k in range(len(anchor_strides)):
base_size = min_sizes[k]
stride = anchor_strides[k]
ctr = ((stride - 1) / 2., (stride - 1) / 2.)
scales = [1., np.sqrt(max_sizes[k] / min_sizes[k])]
ratios = [1.]
for r in anchor_ratios[k]:
ratios += [1 / r, r] # 4 or 6 ratio
anchor_generator = AnchorGenerator(
base_size, scales, ratios, scale_major=False, ctr=ctr)
indices = list(range(len(ratios)))
indices.insert(1, len(indices))
anchor_generator.base_anchors = torch.index_select(
anchor_generator.base_anchors, 0, torch.LongTensor(indices))
self.anchor_generators.append(anchor_generator)
self.target_means = target_means
self.target_stds = target_stds
self.use_sigmoid_cls = False
self.cls_focal_loss = False
self.fp16_enabled = False
def init_weights(self):
for m in self.modules():
if isinstance(m, nn.Conv2d):
xavier_init(m, distribution='uniform', bias=0)
def forward(self, feats):
cls_scores = []
bbox_preds = []
for feat, reg_conv, cls_conv in zip(feats, self.reg_convs,
self.cls_convs):
cls_scores.append(cls_conv(feat))
bbox_preds.append(reg_conv(feat))
return cls_scores, bbox_preds
def loss_single(self, cls_score, bbox_pred, labels, label_weights,
bbox_targets, bbox_weights, num_total_samples, cfg):
loss_cls = GHM_Class_loss(
cls_score,
labels,
label_weights)
loss_bbox = GHM_Reg_loss(
bbox_pred,
bbox_targets,
bbox_weights,
avg_factor=None)
return loss_cls[None], loss_bbox
def loss(self,
cls_scores,
bbox_preds,
gt_bboxes,
gt_labels,
img_metas,
cfg,
gt_bboxes_ignore=None):
featmap_sizes = [featmap.size()[-2:] for featmap in cls_scores]
assert len(featmap_sizes) == len(self.anchor_generators)
anchor_list, valid_flag_list = self.get_anchors(
featmap_sizes, img_metas)
cls_reg_targets = anchor_target(
anchor_list,
valid_flag_list,
gt_bboxes,
img_metas,
self.target_means,
self.target_stds,
cfg,
gt_bboxes_ignore_list=gt_bboxes_ignore,
gt_labels_list=gt_labels,
label_channels=1,
sampling=False,
unmap_outputs=False)
if cls_reg_targets is None:
return None
(labels_list, label_weights_list, bbox_targets_list, bbox_weights_list,
num_total_pos, num_total_neg) = cls_reg_targets
num_images = len(img_metas)
all_cls_scores = torch.cat([
s.permute(0, 2, 3, 1).reshape(
num_images, -1, self.cls_out_channels) for s in cls_scores
], 1)
all_labels = torch.cat(labels_list, -1).view(num_images, -1)
all_label_weights = torch.cat(label_weights_list,
-1).view(num_images, -1)
all_bbox_preds = torch.cat([
b.permute(0, 2, 3, 1).reshape(num_images, -1, 4)
for b in bbox_preds
], -2)
all_bbox_targets = torch.cat(bbox_targets_list,
-2).view(num_images, -1, 4)
all_bbox_weights = torch.cat(bbox_weights_list,
-2).view(num_images, -1, 4)
losses_cls, losses_bbox = multi_apply(
self.loss_single,
all_cls_scores,
all_bbox_preds,
all_labels,
all_label_weights,
all_bbox_targets,
all_bbox_weights,
num_total_samples=num_total_pos,
cfg=cfg)
return dict(loss_cls=losses_cls, loss_bbox=losses_bbox)