#readme.md

VGG16

VGG16 is one of the DNN models for image classification. The original paper of this network is: Very Deep Convolutional Networks for Large-Scale Image Recognition.

Usage

This gist implements an VGG16 service in Owl, and provides simple interfaces to use. Here is an example:

#zoo "f5409c44d6444921a8ceec00e33c42c4"

let img = "/path/to/your/image.png";;
let labels = Vgg16.infer img;;
let labels_json   = Vgg16.to_json ~top:5 labels;;
let labels_tuples = Vgg16.to_tuples labels;;

Note that you need to download the weight file "vgg16_owl.network" from here to current directory before running the code.

The infer function takes image path as input. The image chould be of any popular formats: jpeg, png, etc. This gist contains an exemplar image for you to use, but feel free to use your own.

The output of this function is a 1x1000 vector. The user can further get human-readable classification results by passing this vector to_json or to_tuples. The output of former function is the top-N inference result as a json string, and the latter's is a list, each element in the form of [class: string; propability: float]. The probability is in range [0, 1].

The top parameter specifies how many top-N results are shown. It is default to be 5.

Prerequisite

This application relies on the tool ImageMagick to manipulate image format conversion and resizing. Please make sure it is installed. For example, on Ubuntu or Debian, you can use command:

sudo apt-get install imagemagick

panda.png

vgg16.ml

#!/usr/bin/env owl

open Owl
open Owl_types
open Neural 
open Neural.S
open Neural.S.Graph

#zoo "51eaf74c65fa14c8c466ecfab2351bbd" (* Imagenet_cls *)
#zoo "86a1748bbc898f2e42538839edba00e1" (* ImageUtils   *)

let make_network img_size =
  input [|img_size;img_size;3|]
  (* block 1 *)
  |> conv2d [|3;3;3;64|] [|1;1|] ~act_typ:Activation.Relu ~padding:SAME
  |> conv2d [|3;3;64;64|] [|1;1|] ~act_typ:Activation.Relu ~padding:SAME
  |> max_pool2d [|2;2|] [|2;2|] ~padding:VALID
  (* block 2 *)
  |> conv2d [|3;3;64;128|] [|1;1|] ~act_typ:Activation.Relu ~padding:SAME
  |> conv2d [|3;3;128;128|] [|1;1|] ~act_typ:Activation.Relu ~padding:SAME
  |> max_pool2d [|2;2|] [|2;2|] ~padding:VALID
  (* block 3 *)
  |> conv2d [|3;3;128;256|] [|1;1|] ~act_typ:Activation.Relu ~padding:SAME
  |> conv2d [|3;3;256;256|] [|1;1|] ~act_typ:Activation.Relu ~padding:SAME
  |> conv2d [|3;3;256;256|] [|1;1|] ~act_typ:Activation.Relu ~padding:SAME
  |> max_pool2d [|2;2|] [|2;2|] ~padding:VALID
  (* block 4 *)
  |> conv2d [|3;3;256;512|] [|1;1|] ~act_typ:Activation.Relu ~padding:SAME
  |> conv2d [|3;3;512;512|] [|1;1|] ~act_typ:Activation.Relu ~padding:SAME
  |> conv2d [|3;3;512;512|] [|1;1|] ~act_typ:Activation.Relu ~padding:SAME
  |> max_pool2d [|2;2|] [|2;2|] ~padding:VALID
  (* block 5 *)
  |> conv2d [|3;3;512;512|] [|1;1|] ~act_typ:Activation.Relu ~padding:SAME
  |> conv2d [|3;3;512;512|] [|1;1|] ~act_typ:Activation.Relu ~padding:SAME
  |> conv2d [|3;3;512;512|] [|1;1|] ~act_typ:Activation.Relu ~padding:SAME
  |> max_pool2d [|2;2|] [|2;2|] ~padding:VALID
  (* classification block *)
  |> flatten
  |> fully_connected ~act_typ:Activation.Relu 4096
  |> fully_connected ~act_typ:Activation.Relu 4096
  |> fully_connected ~act_typ:Activation.(Softmax 1) 1000
  |> get_network


let weight_file = Owl_zoo_path.extend_zoo_path "vgg16_owl.network"

let infer img_name = 
  let nn = Graph.load weight_file in 
  let filename = String.split_on_char '/' img_name |> List.rev |> List.hd in
  let prefix   = Filename.remove_extension filename in
  let tmp_img  = Filename.temp_file prefix ".ppm" in 
  let _ = Sys.command ("convert -resize 224x224\\! " ^ img_name ^ " " ^ tmp_img) in
  let img_ppm  = ImageUtils.(load_ppm tmp_img |> extend_dim |> preprocess) in 
  Graph.model nn img_ppm

let to_tuples ?(top=5) label = 
  Imagenet_cls.to_tuples ~top label

let to_json?(top=5) label = 
    Imagenet_cls.to_json ~top label

let test () =
  let example = Owl_zoo_path.extend_zoo_path "panda.png" in
  infer example |> to_json |> Printf.printf "%s\n"

vgg16_owl.network.txt

Please download the weight file "vgg16_owl.network" at: https://drive.google.com/file/d/1tMK6gZsjdD4oT0blWtsZiicliWpploOB/view?usp=sharing