1
以下は、事前に訓練されたモデルを訓練するために使用されるtrain.Prototxtファイルです。caffeとpycaffeによって報告された精度は異なります
name: "TempWLDNET"
layer {
name: "data"
type: "ImageData"
top: "data"
top: "label"
include {
phase: TRAIN
}
transform_param {
mirror: true
crop_size: 224
mean_file: "mean.binaryproto"
}
image_data_param {
source: "train.txt"
batch_size: 25
new_height: 256
new_width: 256
}
}
layer {
name: "data"
type: "ImageData"
top: "data"
top: "label"
include {
phase: TEST
}
transform_param {
mirror: false
crop_size: 224
mean_file: "painmean.binaryproto"
}
image_data_param {
source: "test.txt"
batch_size: 25
new_height: 256
new_width: 256
}
}
layer {
name: "conv1"
type: "Convolution"
bottom: "data"
top: "conv1"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 96
kernel_size: 7
stride: 2
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layer {
name: "relu1"
type: "ReLU"
bottom: "conv1"
top: "conv1"
}
layer {
name: "norm1"
type: "LRN"
bottom: "conv1"
top: "norm1"
lrn_param {
local_size: 5
alpha: 0.0005
beta: 0.75
}
}
layer {
name: "pool1"
type: "Pooling"
bottom: "norm1"
top: "pool1"
pooling_param {
pool: MAX
kernel_size: 3
stride: 3
}
}
layer {
name: "conv2"
type: "Convolution"
bottom: "pool1"
top: "conv2"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 256
pad: 2
kernel_size: 5
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 1
}
}
}
layer {
name: "relu2"
type: "ReLU"
bottom: "conv2"
top: "conv2"
}
layer {
name: "pool2"
type: "Pooling"
bottom: "conv2"
top: "pool2"
pooling_param {
pool: MAX
kernel_size: 2
stride: 2
}
}
layer {
name: "conv3"
type: "Convolution"
bottom: "pool2"
top: "conv3"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 512
pad: 1
kernel_size: 3
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layer {
name: "relu3"
type: "ReLU"
bottom: "conv3"
top: "conv3"
}
layer {
name: "conv4"
type: "Convolution"
bottom: "conv3"
top: "conv4"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 512
pad: 1
kernel_size: 3
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 1
}
}
}
layer {
name: "relu4"
type: "ReLU"
bottom: "conv4"
top: "conv4"
}
layer {
name: "conv5"
type: "Convolution"
bottom: "conv4"
top: "conv5"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 512
pad: 1
kernel_size: 3
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 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: 3
stride: 3
}
}
layer {
name: "fc6"
type: "InnerProduct"
bottom: "pool5"
top: "fc6"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
inner_product_param {
num_output: 4048
weight_filler {
type: "gaussian"
std: 0.005
}
bias_filler {
type: "constant"
value: 1
}
}
}
layer {
name: "relu6"
type: "ReLU"
bottom: "fc6"
top: "fc6"
}
layer {
name: "drop6"
type: "Dropout"
bottom: "fc6"
top: "fc6"
dropout_param {
dropout_ratio: 0.5
}
}
layer {
name: "fc7"
type: "InnerProduct"
bottom: "fc6"
top: "fc7"
# Note that lr_mult can be set to 0 to disable any fine-tuning of this, and any other, layer
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
inner_product_param {
num_output: 4048
weight_filler {
type: "gaussian"
std: 0.005
}
bias_filler {
type: "constant"
value: 1
}
}
}
layer {
name: "relu7"
type: "ReLU"
bottom: "fc7"
top: "fc7"
}
layer {
name: "drop7"
type: "Dropout"
bottom: "fc7"
top: "fc7"
dropout_param {
dropout_ratio: 0.5
}
}
layer {
name: "fc8_temp"
type: "InnerProduct"
bottom: "fc7"
top: "fc8_temp"
# lr_mult is set to higher than for other layers, because this layer is starting from random while the others are already trained
param {
lr_mult: 10
decay_mult: 1
}
param {
lr_mult: 20
decay_mult: 0
}
inner_product_param {
num_output: 16
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layer {
name: "accuracy"
type: "Accuracy"
bottom: "fc8_temp"
bottom: "label"
top: "accuracy"
include {
phase: TEST
}
}
layer {
name: "loss"
type: "SoftmaxWithLoss"
bottom: "fc8_temp"
bottom: "label"
top: "loss"
}
上記のprototxtファイルを使用すると、トレーニング終了時のテストセットで報告された精度は92％です。詳細については、How to evaluate the accuracy and loss of a trained model is good or not in caffe?
を参照してください。私はモデルのスナップショットを13000回の反復で使用し、以下のPythonスクリプトを使用して、混同行列を作成しようとしました。報告される精度は74％です。
#!/usr/bin/python
# -*- coding: utf-8 -*-
import sys
import caffe
import numpy as np
import argparse
from collections import defaultdict
TRAIN_DATA_ROOT='/Images/test/'
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument('--proto', type=str, required=True)
parser.add_argument('--model', type=str, required=True)
parser.add_argument('--meanfile', type=str, required=True)
parser.add_argument('--labelfile', type=str, required=True)
args = parser.parse_args()
proto_data = open(args.meanfile, 'rb').read()
a = caffe.io.caffe_pb2.BlobProto.FromString(proto_data)
mean = caffe.io.blobproto_to_array(a)[0]
caffe.set_mode_gpu()
count = 0
correct = 0
matrix = defaultdict(int) # (real,pred) -> int
labels_set = set()
net = caffe.Net(args.proto, args.model, caffe.TEST)
# load input and configure preprocessing
transformer = caffe.io.Transformer({'data': net.blobs['data'].data.shape})
transformer.set_mean('data', mean)
transformer.set_transpose('data', (2,0,1))
transformer.set_channel_swap('data', (2,1,0))
transformer.set_raw_scale('data', 1)
#note we can change the batch size on-the-fly
#since we classify only one image, we change batch size from 10 to 1
net.blobs['data'].reshape(1,3,224,224)
#load the image in the data layer
f = open(args.labelfile, "r")
for line in f.readlines():
parts = line.split()
example_image = parts[0]
label = int(parts[1])
im = caffe.io.load_image(TRAIN_DATA_ROOT + example_image)
print(im.shape)
net.blobs['data'].data[...] = transformer.preprocess('data', im)
out = net.forward()
plabel = int(out['prob'][0].argmax(axis=0))
count += 1
iscorrect = label == plabel
correct += (1 if iscorrect else 0)
matrix[(label, plabel)] += 1
labels_set.update([label, plabel])
if not iscorrect:
print("\rError: expected %i but predicted %i" \
% (label, plabel))
sys.stdout.write("\rAccuracy: %.1f%%" % (100.*correct/count))
sys.stdout.flush()
print(", %i/%i corrects" % (correct, count))
print ("")
print ("Confusion matrix:")
print ("(r , p) | count")
for l in labels_set:
for pl in labels_set:
print ("(%i , %i) | %i" % (l, pl, matrix[(l,pl)]))
私はdeploy.protxtを使用しています
name: "CaffeNet"
input: "data"
input_shape {
dim: 1
dim: 3
dim: 224
dim: 224
}
layers {
name: "conv1"
type: CONVOLUTION
bottom: "data"
top: "conv1"
blobs_lr: 1
weight_decay: 1
blobs_lr: 2
weight_decay: 0
convolution_param {
num_output: 96
kernel_size: 7
stride: 2
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layers {
name: "relu1"
type: RELU
bottom: "conv1"
top: "conv1"
}
layers {
name: "norm1"
type: LRN
bottom: "conv1"
top: "norm1"
lrn_param {
local_size: 5
alpha: 0.0005
beta: 0.75
}
}
layers {
name: "pool1"
type: POOLING
bottom: "norm1"
top: "pool1"
pooling_param {
pool: MAX
kernel_size: 3
stride: 3
}
}
layers {
name: "conv2"
type: CONVOLUTION
bottom: "pool1"
top: "conv2"
blobs_lr: 1
weight_decay: 1
blobs_lr: 2
weight_decay: 0
convolution_param {
num_output: 256
pad: 2
kernel_size: 5
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 1
}
}
}
layers {
name: "relu2"
type: RELU
bottom: "conv2"
top: "conv2"
}
layers {
name: "pool2"
type: POOLING
bottom: "conv2"
top: "pool2"
pooling_param {
pool: MAX
kernel_size: 2
stride: 2
}
}
layers {
name: "conv3"
type: CONVOLUTION
bottom: "pool2"
top: "conv3"
blobs_lr: 1
weight_decay: 1
blobs_lr: 2
weight_decay: 0
convolution_param {
num_output: 512
pad: 1
kernel_size: 3
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layers {
name: "relu3"
type: RELU
bottom: "conv3"
top: "conv3"
}
layers {
name: "conv4"
type: CONVOLUTION
bottom: "conv3"
top: "conv4"
blobs_lr: 1
weight_decay: 1
blobs_lr: 2
weight_decay: 0
convolution_param {
num_output: 512
pad: 1
kernel_size: 3
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 1
}
}
}
layers {
name: "relu4"
type: RELU
bottom: "conv4"
top: "conv4"
}
layers {
name: "conv5"
type: CONVOLUTION
bottom: "conv4"
top: "conv5"
blobs_lr: 1
weight_decay: 1
blobs_lr: 2
weight_decay: 0
convolution_param {
num_output: 512
pad: 1
kernel_size: 3
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layers {
name: "relu5"
type: RELU
bottom: "conv5"
top: "conv5"
}
layers {
name: "pool5"
type: POOLING
bottom: "conv5"
top: "pool5"
pooling_param {
pool: MAX
kernel_size: 3
stride: 3
}
}
layers {
name: "fc6"
type: INNER_PRODUCT
bottom: "pool5"
top: "fc6"
blobs_lr: 1
weight_decay: 1
blobs_lr: 2
weight_decay: 0
inner_product_param {
num_output: 4048
weight_filler {
type: "gaussian"
std: 0.005
}
bias_filler {
type: "constant"
value: 1
}
}
}
layers {
name: "relu6"
type: RELU
bottom: "fc6"
top: "fc6"
}
layers {
name: "drop6"
type: DROPOUT
bottom: "fc6"
top: "fc6"
dropout_param {
dropout_ratio: 0.5
}
}
layers {
name: "fc7"
type: INNER_PRODUCT
bottom: "fc6"
top: "fc7"
# Note that blobs_lr can be set to 0 to disable any fine-tuning of this, and any other, layers
blobs_lr: 1
weight_decay: 1
blobs_lr: 2
weight_decay: 0
inner_product_param {
num_output: 4048
weight_filler {
type: "gaussian"
std: 0.005
}
bias_filler {
type: "constant"
value: 1
}
}
}
layers {
name: "relu7"
type: RELU
bottom: "fc7"
top: "fc7"
}
layers {
name: "drop7"
type: DROPOUT
bottom: "fc7"
top: "fc7"
dropout_param {
dropout_ratio: 0.5
}
}
layers {
name: "fc8_temp"
type: INNER_PRODUCT
bottom: "fc7"
top: "fc8_temp"
# blobs_lr is set to higher than for other layers, because this layers is starting from random while the others are already trained
blobs_lr: 10
weight_decay: 1
blobs_lr: 20
weight_decay: 0
inner_product_param {
num_output: 16
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layers {
name: "prob"
type: SOFTMAX
bottom: "fc8_temp"
top: "prob"
}
スクリプトを実行するために使用するコマンドは、私が使用しているときの精度に差があるのはなぜ私の疑問がある
python confusion.py --proto deploy.prototxt --model models/model_iter_13000.caffemodel --meanfile mean.binaryproto --labelfile NamesTest.txt
です同じモデルと同じテストセット。何か間違っているのですか？前もって感謝します。
お返事ありがとうございます。 mean.binaryprotoとpainmean.binaryprotoは同じファイルです。それを256x256にスケーリングした後にそれを切り取る方法を知ることができますか？ – Venky
@Venky classifier.pyの例 – Shai
@ Shai、最近私はcaffeを学び始めました。 input_shape {dim：1 dim：3dim：224dim：224}とは何を意味するのでしょうか？ 256x256への入力をどのようにスケーリングし、それを224x224に切り捨てるのか教えてください。私はコードの変更を意味します。以下のコードが実際にはnet.blobs ['data']。data [...]と混同しています。 ] = transformer.preprocess（ 'data'、im）およびnet.blobs ['data']。reshape（1,3,224,224） – Venky