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
|
.\"Text automatically generated by txt2man
.TH CAFFE 1 "10 August 2016" "" ""
.SH NAME
\fBcaffe \fP- command line brew for Caffe
.SH SYNOPSIS
.nf
.fam C
\fBcaffe\fP <COMMAND> <FLAGS>
.fam T
.fi
.fam T
.fi
.SH DESCRIPTION
Caffe is a deep learning framework made with expression, speed,
and modularity in mind. It is developed by the Berkeley Vision
and Learning Center (BVLC) and community contributors.
.SH COMMANDS
.TP
.B
train
train or finetune a model
.TP
.B
test
score a model
.TP
.B
device_query
show GPU diagnostic information
.TP
.B
time
benchmark model execution time
.SH FREQUENTLY USED FLAGS
.TP
.B
\fB-gpu\fP
(Optional; run in GPU mode on given device IDs separated by ','.
Use '\fB-gpu\fP all' to run on all available GPUs. The effective
training batch size is multiplied by the number of devices.)
type: string default: ""
.TP
.B
\fB-iterations\fP
(The number of iterations to run.) type: int32 default: 50
.TP
.B
\fB-level\fP
(Optional; network level.) type: int32 default: 0
.TP
.B
\fB-model\fP
(The model definition protocol buffer text file..) type: string
default: ""
.TP
.B
\fB-phase\fP
(Optional; network phase (TRAIN or TEST). Only used for 'time'.)
type: string default: ""
.TP
.B
\fB-sighup_effect\fP
(Optional; action to take when a SIGHUP signal is received:
snapshot, stop or none.) type: string default: "snapshot"
.TP
.B
\fB-sigint_effect\fP
(Optional; action to take when a SIGINT signal is received:
snapshot, stop or none.) type: string default: "stop"
.TP
.B
\fB-snapshot\fP
(Optional; the snapshot solver state to resume training.)
type: string default: ""
.TP
.B
\fB-solver\fP
(The solver definition protocol buffer text file.) type: string
default: ""
.TP
.B
\fB-stage\fP
(Optional; network stages (not to be confused with phase), separated
by ','.) type: string default: ""
.TP
.B
\fB-weights\fP
(Optional; the pretrained weights to initialize finetuning,
separated by ','. Cannot be set simultaneously with snapshot.)
type: string default: ""
.TP
.B
\fB-help\fP
Show complete help messages.
.SH OTHER CAFFE UTILITIES
Apart from the "\fBcaffe\fP" command line utility, there are also some utilities
available, run them with "\fB-h\fP" or "\fB--help\fP" argument to see corresponding help.
.IP \(bu 3
convert_imageset
.IP \(bu 3
convert_cifar_data
.IP \(bu 3
compute_image_mean
.IP \(bu 3
convert_mnist_siamese_data
.IP \(bu 3
upgrade_net_proto_binary
.IP \(bu 3
extract_features
.IP \(bu 3
upgrade_solver_proto_text
.IP \(bu 3
classification
.IP \(bu 3
upgrade_net_proto_text
.IP \(bu 3
convert_mnist_data
.SH EXAMPLES
Train a new Network
.PP
.nf
.fam C
$ caffe train -solver solver.prototxt
.fam T
.fi
Resume training a network from a snapshot
.PP
.nf
.fam C
$ caffe train -solver solver.prototxt -snapshot bvlc_alexnet.solverstate
.fam T
.fi
Fine-tune a network
.PP
.nf
.fam C
$ caffe train -solver solver.prototxt -weights pre_trained.caffemodel
.fam T
.fi
Test (evaluate) a trained model for 100 iterations, on GPU 0
.PP
.nf
.fam C
$ caffe test -model train_val.prototxt -weights bvlc_alexnet.caffemodel -gpu 0 -iterations 100
.fam T
.fi
Run a benchmark against AlexNet on GPU 0
.PP
.nf
.fam C
$ caffe time -model deploy.prototxt -gpu 0
.fam T
.fi
Check CUDA device availability of GPU 0
.PP
.nf
.fam C
$ caffe device_query -gpu 0
.fam T
.fi
.SH HOMEPAGE
http://caffe.berkeleyvision.org
.SH BUGS
https://github.com/BVLC/\fBcaffe\fP/issues
.SH AUTHOR
This manpage is written by Zhou Mo <cdluminate@gmail.com> with the help of txt2man for Debian
according to program's help message.
|
