glamp

from sklearn.datasets import load_iris
from sklearn.ensemble import RandomForestClassifier
import pandas as pd
import numpy as np

iris = load_iris()
df = pd.DataFrame(iris.data, columns=iris.feature_names)
df['is_train'] = np.random.uniform(0, 1, len(df)) <= .75
df['species'] = pd.Factor(iris.target, iris.target_names)
df.head()

abikoushi

junpenglao

twiecki

import dask
import dask.array as da
import dask.dataframe as dd
import sparse

@dask.delayed(pure=True)
def corr_on_chunked(chunk1, chunk2, corr_thresh=0.9):
    return sparse.COO.from_numpy((np.dot(chunk1, chunk2.T) > corr_thresh))
        
def chunked_corr_sparse_dask(data, chunksize=5000, corr_thresh=0.9):

tanutarou

danstowell

N <- 15
M <- 12
K <-  3
Wconc <- 100
Hconc <- 5
Winit <- structure(c(.056,.111,.056,.111,.056,.111,.056,.111,.056,.111,.056,.111,.111,.056,.111,.056,.111,.056,.111,.056,.111,.056,.111,.056,.063,.063,.125,.063,.063,.125,.063,.063,.125,.063,.063,.125

), .Dim=c(12,3))
X <- structure(c(.032,.032,.091,.032,.157,.264,.157,.139,.486,.257,.709,.934,.036,.036,.08,.036,.255,.192,.255,.1,.638,.171,1.228,.62,.039,.039,.226,.039,.119,.521,.119,.128,.809,.227,.475,1.537,.047,.047,.212,.047,.113,.589,.113,.242,.717,.458,.407,2.002,.02,.02,.058,.02,.07,.189,.07,.107,.251,.203,.293,.705,.033,.033,.062,.033,.128,.268,.128,.208,.325,.403,.549,1.142,.039,.039,.245,.039,.144,.516,.144,.081,.923,.128,.609,1.39,.05,.05,.189,.05,.326,.383,.326,.091,1.077,.136,1.55,1.024,.039,.039,.245,.039,.106,.548,.106,.112,.847,.194,.407,1.565,.023,.023,.134,.023,.072,.301,.072,.065,.486,.112,.291,.862,.04,.04,.145,.04,.205,.352,.205,.131,.726,.233,.938,1.107,.028,.028,.173,.028,.076,.397,.076,.093,.593,.164,.287,1.163,.028,.028,.058,.028,.14,.1

rezoo

Caffe tutorial

この文章ではCNN実装であるCaffeを用いて，特徴ベクトルの抽出やパラメータの学習を行うための方法について説明する．

Caffeでサポートされている機能

以下の作業を行いたいのであれば，Caffeを用いることが望ましい．

• CNNを利用した画像の多クラス分類
• CNNによる特徴ベクトルの抽出
• CNNの転移学習
• Stacked Auto Encoder

fonnesbeck

twiecki

atksh

import warnings
warnings.filterwarnings('ignore')

import pymc3 as pm
from pymc3.distributions.dist_math import bound
import theano.tensor as tt
import theano
import numpy as np
np.random.seed(seed=32)