Caffe accuracy.layer accuracy error when I revise the accuracy - machine-learning

Question

Caffe accuracy.layer accuracy error when I revise the accuracy for my regression project :
the accuracy.layer code to modify accuracy:
for (int i = 0; i < outer_num_; ++i)
for (int j = 0; j < inner_num_; ++j) {
Distance = sqrt((bottom_data[i * dim + j] - bottom_label[i * inner_num_ + j])*(bottom_data[i * dim + j] - bottom_label[i * inner_num_ + j]));
if (Distance <= 10) {
++accuracy;
}
}
}
but the result is:
I1008 22:14:37.701171 102764 caffe.cpp:286] Loss: 70993.9
I1008 22:14:37.701171 102764 caffe.cpp:298] accuracy = -1.#IND
here is my net.prototxt:
layer {
name: "framert"
type: "HDF5Data"
top: "data"
top: "label"
include {
phase: TRAIN
}
hdf5_data_param {
source: "G:/lab-zhang/caffe-windows/data/csv/train_data_list.txt"
batch_size: 10
}
}
layer {
name: "inner1"
type: "InnerProduct"
bottom: "data"
top: "inner1"
param {
lr_mult: 1
decay_mult: 1.5
}
param {
lr_mult: 2
decay_mult: 0
}
inner_product_param {
num_output: 50
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "inner2"
type: "InnerProduct"
bottom: "inner1"
top: "inner2"
param {
lr_mult: 1
decay_mult: 1.0
}
param {
lr_mult: 2
decay_mult: 0
}
inner_product_param {
num_output: 1
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu1"
type: "ReLU"
bottom: "inner2"
top: "inner2"
relu_param {
engine: CAFFE
}
}
layer {
name: "accuracy"
type: "Accuracy"
bottom: "inner2"
bottom: "label"
top: "accuracy"
include {
phase: TEST
}
}
layer {
name: "loss"
type: "EuclideanLoss"
bottom: "inner2"
bottom: "label"
top: "loss"
}
What is the reason for the wrong result: accuracy:-1.#IND ?
here is my net.prototxt:
layer {
name: "framert"
type: "HDF5Data"
top: "data"
top: "label"
include {
phase: TRAIN
}
hdf5_data_param {
source: "G:/lab-zhang/caffe-windows/data/csv/train_data_list.txt"
batch_size: 10
}
}
layer {
name: "inner1"
type: "InnerProduct"
bottom: "data"
top: "inner1"
param {
lr_mult: 1
decay_mult: 1.5
}
param {
lr_mult: 2
decay_mult: 0
}
inner_product_param {
num_output: 50
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "inner2"
type: "InnerProduct"
bottom: "inner1"
top: "inner2"
param {
lr_mult: 1
decay_mult: 1.0
}
param {
lr_mult: 2
decay_mult: 0
}
inner_product_param {
num_output: 1
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu1"
type: "ReLU"
bottom: "inner2"
top: "inner2"
relu_param {
engine: CAFFE
}
}
layer {
name: "accuracy"
type: "Accuracy"
bottom: "inner2"
bottom: "label"
top: "accuracy"
include {
phase: TEST
}
}
layer {
name: "loss"
type: "EuclideanLoss"
bottom: "inner2"
bottom: "label"
top: "loss"
}

Answer 1

The accuracy you get: -1.#IND means the value your code computes is not a number (NaN).
Why you get NaN is unclear from the code you posted. I suspect you changed too much of the accuracy layer code and introduced a bug that led to Nan.
Make sure you do not forget to update count and that you make sure you update top[0]->mutable_cpu_data()[0] with the computed accuracy.
In general, it is best not to override existing layers, but rather write new ones with the desired functionality.
When writing a new layer, please follow the guidelines in caffe wiki and in this git issue. Specifically, write a test for your layer!