vinhkhuc

import java.util.ArrayList;
import java.util.Collections;
import java.util.List;

/**
  OUTPUT:
  + Sum of 10000 double numbers
  For the 1st shuffling, sum = 4999499.9999999935
  For the 2nd shuffling, sum = 4999499.999999965
  

vinhkhuc

"""
Code to replicate Ron Kohavi's cross-validation experiment on the Iris data set.
"""
from sklearn import datasets, svm
from sklearn.cross_validation import cross_val_score, KFold, LeavePOut
import matplotlib.pyplot as plt

output_file = "cross-validation-experiment-iris.png"

iris = datasets.load_iris()

vinhkhuc

from theano import tensor as T, function
x  = T.dscalar('x')
y  = x ** 2
dy = T.grad(cost=y, wrt=x) # Preparing symbolic gradient
df  = function(inputs=[x], outputs=dy)
print(df(4)) # Output: 8

vinhkhuc

# Implementation of a simple MLP network with one hidden layer. Tested on the iris data set.
# Requires: numpy, sklearn, theano

# NOTE: In order to make the code simple, we rewrite x * W_1 + b_1 = x' * W_1' 
# where x' = [x | 1] and W_1' is the matrix W_1 appended with a new row with elements b_1's. 
# Similarly, for h * W_2 + b_2
import theano
from theano import tensor as T
import numpy as np
from sklearn import datasets

vinhkhuc

"""
Vanilla Char-RNN using TensorFlow by Vinh Khuc (@knvinh).
Adapted from Karpathy's min-char-rnn.py
https://gist.github.com/karpathy/d4dee566867f8291f086
Requires tensorflow>=1.0
BSD License
"""
import random
import numpy as np
import tensorflow as tf

vinhkhuc

# Implementation of a simple MLP network with one hidden layer. Tested on the iris data set.
# Requires: numpy, sklearn>=0.18.1, tensorflow>=1.0

# NOTE: In order to make the code simple, we rewrite x * W_1 + b_1 = x' * W_1'
# where x' = [x | 1] and W_1' is the matrix W_1 appended with a new row with elements b_1's.
# Similarly, for h * W_2 + b_2
import tensorflow as tf
import numpy as np
from sklearn import datasets
from sklearn.model_selection import train_test_split

vinhkhuc

vinhkhuc

0.0061718537	-0.0024363345	0.058415722	-0.019220708	-0.051817734	-0.042351343	-0.020908065	0.005183715	-0.013680814	-0.04262279	0.009524972	-0.02140069	0.023697063	0.043885317	0.034181613	-0.0083187	-0.06722789	0.013031735	0.007751275	-0.08180988	0.04614297	-0.06278132	-0.025033472	-0.043717187	-0.020316828	0.050261766	0.060791086	0.05705632	-0.012153195	0.03464042	-0.08152679	-0.058986522	0.0035639652	0.024647668	-0.059734866	0.015541221	-0.0065737423	0.020316336	-0.0148979565	-0.026047753	0.028355088	-0.026079465	-0.002358603	-0.0067062867	-0.009114266	0.0010216745	0.026439784	0.07460034	0.044666305	0.037209775	0.074898235	0.021767966	-0.052332643	-0.022787187	0.044321213	0.009379231	-0.026886253	0.034125276	-0.06483722	-0.01966572	-0.040497903	-0.0055085467	0.0017649964	0.021614494	-0.009420735	0.05497118	0.010820785	-0.01948987	-0.012267729	-0.010204952	-0.08628109	-0.083253555	-0.004771075	-0.021505628	0.083285354	0.026299926	0.057961594	0.02451058	0.023163367	0.018347874	-0.04912325	-0.03191776	-0.03019

vinhkhuc

{
  "embeddings": [
    {
      "tensorName": "QC Question Embeddings Using TF Sentence Embedding Encoder Large 3",
      "tensorShape": [
        5451,
        300
      ],
      "tensorPath": "https://gist.githubusercontent.com/vinhkhuc/3806dc8ef988d3c10ecfa6310de1943f/raw/738c1d07563612df2662a0aedc931a89cb64458b/qc-question-embeddings.tsv",
      "metadataPath": "https://gist.githubusercontent.com/vinhkhuc/1f9fc9ad322c152ebe607e9bb5d7da55/raw/1518504769d8989f017caa171313991ae43f7e75/qc-embed-metadata.tsv"

vinhkhuc

https://projector.tensorflow.org/?config=https://gist.githubusercontent.com/vinhkhuc/5ca429d4c38900d9eb24a8c245593c75/raw/2f9fe9fb9ae824d030ff7dd6eb941e6dfef6544a/qc-question-embeddings-projector-config.json