jzstark

# DATA_URL = 'http://download.tensorflow.org/models/image/imagenet/inception-2015-12-05.tgz'

def create_graph():
  """Creates a graph from saved GraphDef file and returns a saver."""
  # Creates graph from saved graph_def.pb.
  with tf.gfile.FastGFile('classify_image_graph_def.pb', 'rb') as f:
    graph_def = tf.GraphDef()
    graph_def.ParseFromString(f.read())
    _ = tf.import_graph_def(graph_def, name='')

jzstark

jzstark

(*A good example*)
let y = Mat.(gaussian 100 1 *$ 10.) in
let h = create "" in 
let _ = set_background_color h 28 28 28 in
let _ = set_title h "Gaussian Data to fit Weibull Dist" in 
let _ = set_xlabel h "Sample Data" in
let _ = set_ylabel h "Theoratical Weibull Distribution invCdf" in
let _ = probplot ~h ~dist:(fun i -> Owl_stats.Cdf.weibull_Pinv i 1. 1.) y in
output h;;

jzstark

(*example*)

(*Add logscale: Add xlogscale and ylogscale properties to a page; 
  The logdata still needs to be precomputed.*)

let x = Mat.logspace (-1.) 2. 10 in 
let y = Mat.map Maths.exp x in
let z = Mat.log10 y in 
let h = create "log4.pdf" in
let _ = scatter ~h ~color:(255,0,0)~marker:"#[0x2299]" ~marker_size:8. x z in

jzstark

jzstark

231	210	156	159	159	159	160	154	166	163	166	167	160	168	171	175	170	166	175	167	169	175	165	168	169	166	177	174	167	169	172	177	175	171	173	172	161	167	175	172	167	177	172	169	169	171	170	180	172	172	171	167	177	178	173	183	166	173	171	175	175	169	173	173	170	166	170	172	164	172	171	171	174	168	170	177	162	169	174	160	165	172	172	174	173	174	174	172	170	171	175	168	182	170	172	166	179	176	172	171	170	175	177	169	170	175	172	173	171	176	173	177	177	187	170	179	177	181	181	177	181	176	183	179	182	188	182	183	184	186	188	182	186	182	183	184	182	191	184	189	187	189	185	185	184	190	192	193	190	187	192	185	185	186	191	188	199	191	197	193	197	202	185	198	196	189	197	198	184	201	201	199	192	205	196	204	201	206	208	208	203	199	217	205	205	204	213	204	210	201	221	205	213	208	218	212	215	211	215	213	218	211	216	218	210	226	217	212	218	220	214	222	222	227	219	222	218	216	221	224	218	214	222	224	214	219	220	222	219	216	228	219	225	225	226	223	225	222	219	227	220	234	207	221	217	227	217	228	220	227	219	232	218	230	221	225	

jzstark

open Owl
open Owl_neural
open Algodiff.S
open Owl_neural_neuron

(*
let test_cnn nn x y =
  for i = 0 to 9 do
    let u = Dense.Ndarray.S.slice [[i]] x in
    Dense.Ndarray.S.reshape u [|3;32;32|]

jzstark

open Owl_neural
open Owl_neural_graph
open Algodiff.S
open Owl_neural_neuron
open Plplot


let x, y = Dataset.load_cifar_train_data 1
let m = Dense.Matrix.S.row_num x
let num_test = 9

jzstark

(* source: http://ai.stanford.edu/~amaas/data/sentiment/ *)

let vocab_file = "imdb.vocab"

let load_file f =
  let ic = open_in f in
  let n = in_channel_length ic in
  let s = Bytes.create n in
  really_input ic s 0 n;
  close_in ic;

jzstark

(*
Example script to generate text from Nietzsche's writings.
At least 20 epochs are required before the generated text
starts sounding coherent.
It is recommended to run this script on GPU, as recurrent
networks are quite computationally intensive.

Adapted from Keras example: lstm_text_generation.py
*)