Skip to content

Instantly share code, notes, and snippets.

View qzhong0605's full-sized avatar

qzhong0605

12 followers · 37 following
  • Beijing, China
View GitHub Profile
@qzhong0605
qzhong0605 / deploy_no_inplace_inception-v3.prototxt
Created November 30, 2018 02:30
name: "Inception-V3(no-inplace)"
input: "data"
input_shape {
dim: 1
dim: 3
dim: 299
dim: 299
}
layer {
name: "conv1_3x3_s2"
@qzhong0605
qzhong0605 / deploy_resnet101-v2.prototxt
Created November 30, 2018 05:51
name: "ResNet-101v2"
input: "data"
input_shape {
dim: 1
dim: 3
dim: 224
dim: 224
}
layer {
@qzhong0605
qzhong0605 / deploy_no_inplace_resnet101-v2.prototxt
Created November 30, 2018 05:55
name: "ResNet-101v2(no-inplace)"
input: "data"
input_shape {
dim: 1
dim: 3
dim: 224
dim: 224
}
layer {
@qzhong0605
qzhong0605 / deploy_no_inplace_resnet152-v2.prototxt
Created November 30, 2018 08:04
name: "ResNet-152v2"
input: "data"
input_shape {
dim: 1
dim: 3
dim: 224
dim: 224
}
layer {
@qzhong0605
qzhong0605 / deploy_resnet152-v2.prototxt
Created November 30, 2018 08:10
name: "ResNet-152v2"
input: "data"
input_shape {
dim: 1
dim: 3
dim: 320
dim: 320
}
layer {
@qzhong0605
qzhong0605 / resnet18_train_val.prototxt
Created December 1, 2018 07:36
name: "ResNet-18"
layer {
name: "data"
type: "Data"
top: "data"
top: "label"
include {
phase: TRAIN
}
transform_param {
@qzhong0605
qzhong0605 / resnet18_no_inplace_deploy.prototxt
Created December 1, 2018 07:43
name: "ResNet-18"
layer {
name: "data"
type: "Input"
top: "data"
input_param {
shape {
dim: 1
dim: 3
dim: 224
@qzhong0605
qzhong0605 / se_resnet50_train_val.prototxt
Created December 1, 2018 08:41
name: "SE_RESNET_50_V1"
layer {
name: "data"
type: "Data"
top: "data"
top: "label"
include {
phase: TRAIN
}
@qzhong0605
qzhong0605 / se_resnet_50_v1_deploy.prototxt
Created December 1, 2018 08:43
name: "SE_RESNET_50_V1"
input: "data"
input_shape {
dim: 1
dim: 3
dim: 225
dim: 225
}
# transform_param {
# scale: 0.017
@qzhong0605
qzhong0605 / se_resnet_50_v1_no_inplace_deploy.prototxt
Created December 1, 2018 15:33
name: "SE_RESNET_50_V1"
input: "data"
input_shape {
dim: 1
dim: 3
dim: 224
dim: 224
}
layer {
name: "conv1"