I have a small network as follow. I have 1400 trained images and 350 test images. I trained from scratch without any pretrained weight. Actually it is same as SSD and just that network architecture is changed. So loss function is also same as SSD.
name: "train"
layer {
name: "data"
type: "AnnotatedData"
top: "data"
top: "label"
include {
phase: TRAIN
}
transform_param {
mirror: false
mean_value: 109.0
mean_value: 110.0
mean_value: 110.0
resize_param {
prob: 1.0
resize_mode: WARP
height: 480
width: 480
interp_mode: LINEAR
interp_mode: AREA
interp_mode: NEAREST
interp_mode: CUBIC
interp_mode: LANCZOS4
height_scale: 480
width_scale: 480
}
emit_constraint {
emit_type: CENTER
}
distort_param {
brightness_prob: 0.5
brightness_delta: 32.0
contrast_prob: 0.5
contrast_lower: 0.5
contrast_upper: 1.5
hue_prob: 0.5
hue_delta: 18.0
saturation_prob: 0.5
saturation_lower: 0.5
saturation_upper: 1.5
random_order_prob: 0.0
}
expand_param {
prob: 0.5
max_expand_ratio: 4.0
}
}
data_param {
source: "/home/coie/data/NumberPlate/lmdb/Nextan_trainval_lmdb"
batch_size: 16
backend: LMDB
}
annotated_data_param {
batch_sampler {
max_sample: 1
max_trials: 1
}
batch_sampler {
sampler {
min_scale: 0.300000011921
max_scale: 1.0
min_aspect_ratio: 0.5
max_aspect_ratio: 2.0
}
sample_constraint {
min_jaccard_overlap: 0.5
}
max_sample: 1
max_trials: 50
}
batch_sampler {
sampler {
min_scale: 0.300000011921
max_scale: 1.0
min_aspect_ratio: 0.5
max_aspect_ratio: 2.0
}
sample_constraint {
min_jaccard_overlap: 0.699999988079
}
max_sample: 1
max_trials: 50
}
batch_sampler {
sampler {
min_scale: 0.300000011921
max_scale: 1.0
min_aspect_ratio: 0.5
max_aspect_ratio: 2.0
}
sample_constraint {
min_jaccard_overlap: 0.899999976158
}
max_sample: 1
max_trials: 50
}
batch_sampler {
sampler {
min_scale: 0.300000011921
max_scale: 1.0
min_aspect_ratio: 0.5
max_aspect_ratio: 2.0
}
sample_constraint {
max_jaccard_overlap: 1.0
}
max_sample: 1
max_trials: 50
}
label_map_file: "/home/coie/data/NumberPlate/labelmap_NumberPlate.prototxt"
}
}
layer {
name: "conv1"
type: "Convolution"
bottom: "data"
top: "conv1"
param {
lr_mult: 1.0
decay_mult: 1.0
}
param {
lr_mult: 2.0
decay_mult: 0.0
}
convolution_param {
num_output: 6
pad: 1
kernel_size: 3
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
value: 0.0
}
}
}
layer {
name: "relu1"
type: "ReLU"
bottom: "conv1"
top: "conv1"
}
layer {
name: "conv2"
type: "Convolution"
bottom: "conv1"
top: "conv2"
param {
lr_mult: 1.0
decay_mult: 1.0
}
param {
lr_mult: 2.0
decay_mult: 0.0
}
convolution_param {
num_output: 6
pad: 1
kernel_size: 3
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
value: 0.0
}
}
}
layer {
name: "relu2"
type: "ReLU"
bottom: "conv2"
top: "conv2"
}
layer {
name: "pool2"
type: "Pooling"
bottom: "conv2"
top: "pool2"
pooling_param {
pool: MAX
kernel_size: 3
stride: 3
}
}
layer {
name: "conv3"
type: "Convolution"
bottom: "pool2"
top: "conv3"
param {
lr_mult: 1.0
decay_mult: 1.0
}
param {
lr_mult: 2.0
decay_mult: 0.0
}
convolution_param {
num_output: 16
pad: 1
kernel_size: 3
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
value: 0.0
}
}
}
layer {
name: "relu3"
type: "ReLU"
bottom: "conv3"
top: "conv3"
}
layer {
name: "pool3"
type: "Pooling"
bottom: "conv3"
top: "pool3"
pooling_param {
pool: MAX
kernel_size: 2
stride: 2
}
}
layer {
name: "conv4"
type: "Convolution"
bottom: "pool3"
top: "conv4"
param {
lr_mult: 1.0
decay_mult: 1.0
}
param {
lr_mult: 2.0
decay_mult: 0.0
}
convolution_param {
num_output: 32
pad: 1
kernel_size: 3
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
value: 0.0
}
}
}
layer {
name: "relu4"
type: "ReLU"
bottom: "conv4"
top: "conv4"
}
layer {
name: "pool4"
type: "Pooling"
bottom: "conv4"
top: "pool4"
pooling_param {
pool: MAX
kernel_size: 2
stride: 2
}
}
layer {
name: "conv5"
type: "Convolution"
bottom: "pool4"
top: "conv5"
param {
lr_mult: 1.0
decay_mult: 1.0
}
param {
lr_mult: 2.0
decay_mult: 0.0
}
convolution_param {
num_output: 64
pad: 1
kernel_size: 3
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
value: 0.0
}
}
}
layer {
name: "relu5"
type: "ReLU"
bottom: "conv5"
top: "conv5"
}
layer {
name: "pool5"
type: "Pooling"
bottom: "conv5"
top: "pool5"
pooling_param {
pool: MAX
kernel_size: 2
stride: 2
}
}
layer {
name: "conv6"
type: "Convolution"
bottom: "pool5"
top: "conv6"
param {
lr_mult: 1.0
decay_mult: 1.0
}
param {
lr_mult: 2.0
decay_mult: 0.0
}
convolution_param {
num_output: 64
pad: 1
kernel_size: 3
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
value: 0.0
}
}
}
layer {
name: "relu6"
type: "ReLU"
bottom: "conv6"
top: "conv6"
}
layer {
name: "pool6"
type: "Pooling"
bottom: "conv6"
top: "pool6"
pooling_param {
pool: MAX
kernel_size: 2
stride: 2
}
}
layer {
name: "conv7"
type: "Convolution"
bottom: "pool6"
top: "conv7"
param {
lr_mult: 1.0
decay_mult: 1.0
}
param {
lr_mult: 2.0
decay_mult: 0.0
}
convolution_param {
num_output: 64
pad: 1
kernel_size: 3
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
value: 0.0
}
}
}
layer {
name: "relu7"
type: "ReLU"
bottom: "conv7"
top: "conv7"
}
layer {
name: "conv6_norm"
type: "Normalize"
bottom: "conv6"
top: "conv6_norm"
norm_param {
across_spatial: false
scale_filler {
type: "constant"
value: 20.0
}
channel_shared: false
}
}
layer {
name: "conv6_norm_mbox_loc"
type: "Convolution"
bottom: "conv6_norm"
top: "conv6_norm_mbox_loc"
param {
lr_mult: 1.0
decay_mult: 1.0
}
param {
lr_mult: 2.0
decay_mult: 0.0
}
convolution_param {
num_output: 12
pad: 1
kernel_size: 3
stride: 1
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
value: 0.0
}
}
}
layer {
name: "conv6_norm_mbox_loc_perm"
type: "Permute"
bottom: "conv6_norm_mbox_loc"
top: "conv6_norm_mbox_loc_perm"
permute_param {
order: 0
order: 2
order: 3
order: 1
}
}
layer {
name: "conv6_norm_mbox_loc_flat"
type: "Flatten"
bottom: "conv6_norm_mbox_loc_perm"
top: "conv6_norm_mbox_loc_flat"
flatten_param {
axis: 1
}
}
layer {
name: "conv6_norm_mbox_conf"
type: "Convolution"
bottom: "conv6_norm"
top: "conv6_norm_mbox_conf"
param {
lr_mult: 1.0
decay_mult: 1.0
}
param {
lr_mult: 2.0
decay_mult: 0.0
}
convolution_param {
num_output: 6
pad: 1
kernel_size: 3
stride: 1
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
value: 0.0
}
}
}
layer {
name: "conv6_norm_mbox_conf_perm"
type: "Permute"
bottom: "conv6_norm_mbox_conf"
top: "conv6_norm_mbox_conf_perm"
permute_param {
order: 0
order: 2
order: 3
order: 1
}
}
layer {
name: "conv6_norm_mbox_conf_flat"
type: "Flatten"
bottom: "conv6_norm_mbox_conf_perm"
top: "conv6_norm_mbox_conf_flat"
flatten_param {
axis: 1
}
}
layer {
name: "conv6_norm_mbox_priorbox"
type: "PriorBox"
bottom: "conv6_norm"
bottom: "data"
top: "conv6_norm_mbox_priorbox"
prior_box_param {
min_size: 20.0
max_size: 40.0
aspect_ratio: 2.0
flip: false
clip: false
variance: 0.10000000149
variance: 0.10000000149
variance: 0.20000000298
variance: 0.20000000298
img_size: 480
step: 24.0
offset: 0.5
}
}
layer {
name: "conv7_norm"
type: "Normalize"
bottom: "conv7"
top: "conv7_norm"
norm_param {
across_spatial: false
scale_filler {
type: "constant"
value: 20.0
}
channel_shared: false
}
}
layer {
name: "conv7_norm_mbox_loc"
type: "Convolution"
bottom: "conv7_norm"
top: "conv7_norm_mbox_loc"
param {
lr_mult: 1.0
decay_mult: 1.0
}
param {
lr_mult: 2.0
decay_mult: 0.0
}
convolution_param {
num_output: 12
pad: 1
kernel_size: 3
stride: 1
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
value: 0.0
}
}
}
layer {
name: "conv7_norm_mbox_loc_perm"
type: "Permute"
bottom: "conv7_norm_mbox_loc"
top: "conv7_norm_mbox_loc_perm"
permute_param {
order: 0
order: 2
order: 3
order: 1
}
}
layer {
name: "conv7_norm_mbox_loc_flat"
type: "Flatten"
bottom: "conv7_norm_mbox_loc_perm"
top: "conv7_norm_mbox_loc_flat"
flatten_param {
axis: 1
}
}
layer {
name: "conv7_norm_mbox_conf"
type: "Convolution"
bottom: "conv7_norm"
top: "conv7_norm_mbox_conf"
param {
lr_mult: 1.0
decay_mult: 1.0
}
param {
lr_mult: 2.0
decay_mult: 0.0
}
convolution_param {
num_output: 6
pad: 1
kernel_size: 3
stride: 1
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
value: 0.0
}
}
}
layer {
name: "conv7_norm_mbox_conf_perm"
type: "Permute"
bottom: "conv7_norm_mbox_conf"
top: "conv7_norm_mbox_conf_perm"
permute_param {
order: 0
order: 2
order: 3
order: 1
}
}
layer {
name: "conv7_norm_mbox_conf_flat"
type: "Flatten"
bottom: "conv7_norm_mbox_conf_perm"
top: "conv7_norm_mbox_conf_flat"
flatten_param {
axis: 1
}
}
layer {
name: "conv7_norm_mbox_priorbox"
type: "PriorBox"
bottom: "conv7_norm"
bottom: "data"
top: "conv7_norm_mbox_priorbox"
prior_box_param {
min_size: 40.0
max_size: 60.0
aspect_ratio: 2.0
flip: false
clip: false
variance: 0.10000000149
variance: 0.10000000149
variance: 0.20000000298
variance: 0.20000000298
img_size: 480
step: 48.0
offset: 0.5
}
}
layer {
name: "mbox_loc"
type: "Concat"
bottom: "conv6_norm_mbox_loc_flat"
bottom: "conv7_norm_mbox_loc_flat"
top: "mbox_loc"
concat_param {
axis: 1
}
}
layer {
name: "mbox_conf"
type: "Concat"
bottom: "conv6_norm_mbox_conf_flat"
bottom: "conv7_norm_mbox_conf_flat"
top: "mbox_conf"
concat_param {
axis: 1
}
}
layer {
name: "mbox_priorbox"
type: "Concat"
bottom: "conv6_norm_mbox_priorbox"
bottom: "conv7_norm_mbox_priorbox"
top: "mbox_priorbox"
concat_param {
axis: 2
}
}
layer {
name: "mbox_loss"
type: "MultiBoxLoss"
bottom: "mbox_loc"
bottom: "mbox_conf"
bottom: "mbox_priorbox"
bottom: "label"
top: "mbox_loss"
include {
phase: TRAIN
}
propagate_down: true
propagate_down: true
propagate_down: false
propagate_down: false
loss_param {
normalization: VALID
}
multibox_loss_param {
loc_loss_type: SMOOTH_L1
conf_loss_type: SOFTMAX
loc_weight: 1.0
num_classes: 2
share_location: true
match_type: PER_PREDICTION
overlap_threshold: 0.5
use_prior_for_matching: true
background_label_id: 1
use_difficult_gt: true
neg_pos_ratio: 3.0
neg_overlap: 0.5
code_type: CENTER_SIZE
ignore_cross_boundary_bbox: false
mining_type: MAX_NEGATIVE
}
}
I trained for about 5000 iterations. Detection_eval is quite ok 82% and no overfitting. But loss is not low around 5. What should I do to reduce the loss? Training loss and iterations, Learning rate and iterations are attached.
