Skip to content

Nimonm/machine-learning-assignment_2

Repository files navigation

machine-learning-assignment_2

Writen by: Julie

A. Visualization of neurons via input optimization

The code is in Ass_A.py, which is changed by the original code named 'myalexnet_forward_newtf.py'. And there are some outputs generated by this experiment with different learning accuracy and gradient update function, all of them are in the file named 'A_gen'.

B. Natural image statistics

The code is in Ass_B.py, and I use the 122th class(king crab, Alaska crab, Alaskan king crab, Alaska king crab, Paralithodes camtschatica) for this section. Due to time constraints, I only ran 1200 steps. So it can not do well enough(I am sorry).

C. Fooling a network

The code is in Ass_C.py, and the example output files are 'C_poodle', 'C_laska'. In each file, every image named by 'number_class-names_probability', 'number' means the image generation order, 'class-name' means which class this image be identified, 'probability' means the probability that the picture is recognized as this class.

D. Classification importance map

The code is in 'Ass_D.py', and the output importance maps are in files named 'D_dog', 'D_poodle', which shows importance maps in each iteration.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages