How could I draw the average loss line in training?

[john1...@gmail.com]

Hello, I got the result of training data after 40.000 iterations by using script

./build/tools/caffe train --solver=solver.prototxt -gpu 0 2>&1 | tee -a my_model.log

Then, I parse the log file and draw it as bellow figure by using script

python ~/caffe/tools/extra/parse_log.py my_model.log .
gnuplot -persist gnuplot_commands

where gnuplot_commands is a file that stores a set of gnuplot commands.

set datafile separator ','
set term x11 0
plot '../my_model.log.train' using 1:4  with line

Finally, I achieved the result as follows:

Training loss

I have a question about the above figure

1. How could I draw the average loss line, instead of loss raw from above script? I want to use average loss to see the line gradually decrease 

2. As shown in the figure, the training loss at some point grows up very larger. Do you guess any reason?

Thank you in advance