Skip to content

Instantly share code, notes, and snippets.

@e96031413
Created October 26, 2019 04:35
Show Gist options
  • Save e96031413/d703c1a6c35fe3923122ce543a29642f to your computer and use it in GitHub Desktop.
Save e96031413/d703c1a6c35fe3923122ce543a29642f to your computer and use it in GitHub Desktop.
YOLOv2 Tiny的cfg檔案(針對單class)
[net]
# Testing
batch=16
subdivisions=16
# Training
# batch=64
# subdivisions=2
width=416
height=416
channels=3
momentum=0.9
decay=0.0005
angle=0
saturation = 1.5
exposure = 1.5
hue=.1
learning_rate=0.001
burn_in=1000
max_batches = 500200
policy=steps
steps=400000,450000
scales=.1,.1
[convolutional]
batch_normalize=1
filters=16
size=3
stride=1
pad=1
activation=leaky
[maxpool]
size=2
stride=2
[convolutional]
batch_normalize=1
filters=32
size=3
stride=1
pad=1
activation=leaky
[maxpool]
size=2
stride=2
[convolutional]
batch_normalize=1
filters=64
size=3
stride=1
pad=1
activation=leaky
[maxpool]
size=2
stride=2
[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=leaky
[maxpool]
size=2
stride=2
[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=leaky
[maxpool]
size=2
stride=2
[convolutional]
batch_normalize=1
filters=512
size=3
stride=1
pad=1
activation=leaky
[maxpool]
size=2
stride=1
[convolutional]
batch_normalize=1
filters=1024
size=3
stride=1
pad=1
activation=leaky
###########
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=leaky
[convolutional]
size=1
stride=1
pad=1
filters=30
activation=linear
[region]
anchors = 0.57273, 0.677385, 1.87446, 2.06253, 3.33843, 5.47434, 7.88282, 3.52778, 9.77052, 9.16828
bias_match=1
classes=1
coords=4
num=5
softmax=1
jitter=.2
rescore=0
object_scale=5
noobject_scale=1
class_scale=1
coord_scale=1
absolute=1
thresh = .6
random=1
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment