!pip install kaggle

Requirement already satisfied: kaggle in /usr/local/lib/python3.6/dist-packages (1.5.6)
Requirement already satisfied: requests in /usr/local/lib/python3.6/dist-packages (from kaggle) (2.21.0)
Requirement already satisfied: python-slugify in /usr/local/lib/python3.6/dist-packages (from kaggle) (4.0.0)
Requirement already satisfied: certifi in /usr/local/lib/python3.6/dist-packages (from kaggle) (2019.9.11)
Requirement already satisfied: six>=1.10 in /usr/local/lib/python3.6/dist-packages (from kaggle) (1.12.0)
Requirement already satisfied: urllib3<1.25,>=1.21.1 in /usr/local/lib/python3.6/dist-packages (from kaggle) (1.24.3)
Requirement already satisfied: tqdm in /usr/local/lib/python3.6/dist-packages (from kaggle) (4.28.1)
Requirement already satisfied: python-dateutil in /usr/local/lib/python3.6/dist-packages (from kaggle) (2.6.1)
Requirement already satisfied: idna<2.9,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests->kaggle) (2.8)
Requirement already satisfied: chardet<3.1.0,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests->kaggle) (3.0.4)
Requirement already satisfied: text-unidecode>=1.3 in /usr/local/lib/python3.6/dist-packages (from python-slugify->kaggle) (1.3)

Here I will train a machine learning and deep learning model to recognize "Kannad" digits.

Dataset has been downloaded from an active kaggle competition.

That I also participated

from google.colab import files

uploaded = files.upload()

for fn in uploaded.keys():
  print('User uploaded file "{name}" with length {length} bytes'.format(
      name=fn, length=len(uploaded[fn])))
  

!mkdir -p ~/.kaggle/ && mv kaggle.json ~/.kaggle/ && chmod 600 ~/.kaggle/kaggle.json

Upload widget is only available when the cell has been executed in the current browser session. Please rerun this cell to enable.

Saving kaggle.json to kaggle.json
User uploaded file "kaggle.json" with length 68 bytes

Listing competitions

!kaggle competitions list

Warning: Looks like you're using an outdated API Version, please consider updating (server 1.5.6 / client 1.5.4)
ref                                            deadline             category            reward  teamCount  userHasEntered  
---------------------------------------------  -------------------  ---------------  ---------  ---------  --------------  
digit-recognizer                               2030-01-01 00:00:00  Getting Started  Knowledge       2337           False  
titanic                                        2030-01-01 00:00:00  Getting Started  Knowledge      16498            True  
house-prices-advanced-regression-techniques    2030-01-01 00:00:00  Getting Started  Knowledge       5325            True  
imagenet-object-localization-challenge         2029-12-31 07:00:00  Research         Knowledge         57           False  
google-quest-challenge                         2020-02-10 23:59:00  Featured           $25,000        143           False  
tensorflow2-question-answering                 2020-01-22 23:59:00  Featured           $50,000        699           False  
data-science-bowl-2019                         2020-01-22 23:59:00  Featured          $160,000       1325            True  
pku-autonomous-driving                         2020-01-21 23:59:00  Featured           $25,000        369           False  
competitive-data-science-predict-future-sales  2019-12-31 23:59:00  Playground           Kudos       4934            True  
ashrae-energy-prediction                       2019-12-19 23:59:00  Featured           $25,000       2621           False  
Kannada-MNIST                                  2019-12-17 23:59:00  Playground       Knowledge        875           False  
bigquery-geotab-intersection-congestion        2019-12-12 23:59:00  Playground           Kudos        398           False  
cat-in-the-dat                                 2019-12-09 23:59:00  Playground            Swag       1173           False  
kaggle-survey-2019                             2019-12-02 23:59:00  Analytics          $30,000          0           False  
nfl-big-data-bowl-2020                         2019-11-27 23:59:00  Featured           $75,000       1915            True  
understanding_cloud_organization               2019-11-18 23:59:00  Research           $10,000       1538           False  
rsna-intracranial-hemorrhage-detection         2019-11-13 23:59:00  Featured           $25,000       1345           False  
3d-object-detection-for-autonomous-vehicles    2019-11-12 23:59:00  Featured           $25,000        549           False  
severstal-steel-defect-detection               2019-10-24 23:59:00  Featured          $120,000       2433           False  
kuzushiji-recognition                          2019-10-14 23:59:00  Playground         $15,000        293           False  

Downloading a dataset

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from fastai import *
from fastai.vision import *
import pandas as pd

!unzip Kannada-MNIST.zip

Archive:  Kannada-MNIST.zip
  inflating: Dig-MNIST.csv           
  inflating: sample_submission.csv   
  inflating: test.csv                
  inflating: train.csv               

df=pd.read_csv('train.csv')

	label	pixel0	pixel1	pixel2	pixel3	pixel4	pixel5	pixel6	pixel7	pixel8	pixel9	pixel10	pixel11	pixel12	pixel13	pixel14	pixel15	pixel16	pixel17	pixel18	pixel19	pixel20	pixel21	pixel22	pixel23	pixel24	pixel25	pixel26	pixel27	pixel28	pixel29	pixel30	pixel31	pixel32	pixel33	pixel34	pixel35	pixel36	pixel37	pixel38	...	pixel744	pixel745	pixel746	pixel747	pixel748	pixel749	pixel750	pixel751	pixel752	pixel753	pixel754	pixel755	pixel756	pixel757	pixel758	pixel759	pixel760	pixel761	pixel762	pixel763	pixel764	pixel765	pixel766	pixel767	pixel768	pixel769	pixel770	pixel771	pixel772	pixel773	pixel774	pixel775	pixel776	pixel777	pixel778	pixel779	pixel780	pixel781	pixel782	pixel783
0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
2	2	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
3	3	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
4	4	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
59995	5	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
59996	6	0	0	0	0	0	0	0	0	0	0	0	0	147	49	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
59997	7	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	20	255	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
59998	8	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
59999	9	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0

60000 rows × 785 columns

df.head(20)

	label	pixel0	pixel1	pixel2	pixel3	pixel4	pixel5	pixel6	pixel7	pixel8	pixel9	pixel10	pixel11	pixel12	pixel13	pixel14	pixel15	pixel16	pixel17	pixel18	pixel19	pixel20	pixel21	pixel22	pixel23	pixel24	pixel25	pixel26	pixel27	pixel28	pixel29	pixel30	pixel31	pixel32	pixel33	pixel34	pixel35	pixel36	pixel37	pixel38	...	pixel744	pixel745	pixel746	pixel747	pixel748	pixel749	pixel750	pixel751	pixel752	pixel753	pixel754	pixel755	pixel756	pixel757	pixel758	pixel759	pixel760	pixel761	pixel762	pixel763	pixel764	pixel765	pixel766	pixel767	pixel768	pixel769	pixel770	pixel771	pixel772	pixel773	pixel774	pixel775	pixel776	pixel777	pixel778	pixel779	pixel780	pixel781	pixel782	pixel783
0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
2	2	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
3	3	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
4	4	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
5	5	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
6	6	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
7	7	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
8	8	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
9	9	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
10	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
11	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
12	2	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
13	3	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
14	4	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
15	5	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
16	6	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
17	7	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
18	8	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
19	9	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0

20 rows × 785 columns

test=pd.read_csv('test.csv')

k0=df.iloc[0].values
k1=df.iloc[1].values
k2=df.iloc[2].values
k3=df.iloc[3].values
k4=df.iloc[4].values
k5=df.iloc[5].values
k6=df.iloc[6].values
k7=df.iloc[7].values
k8=df.iloc[8].values
k9=df.iloc[9].values

zero=k0[1:].reshape((28,28))
one=k1[1:].reshape((28,28))
two=k2[1:].reshape((28,28))
three=k3[1:].reshape((28,28))
four=k4[1:].reshape((28,28))
five=k5[1:].reshape((28,28))
six=k6[1:].reshape((28,28))
seven=k7[1:].reshape((28,28))
eight=k8[1:].reshape((28,28))
nine=k9[1:].reshape((28,28))

number=['zero','one','two','three','four','five','six','seven','eight','nine']

plt.figure(1,figsize=(20,20))
#for i in range(431,440):
plt.subplot(431)
plt.imshow(zero);
plt.title('zero')
plt.colorbar()
plt.subplot(432)
plt.imshow(one);
plt.title('one')
plt.colorbar()
plt.subplot(433)
plt.imshow(two);
plt.title('two')
plt.colorbar()
plt.subplot(434)
plt.imshow(three);
plt.title('three')
plt.colorbar()
plt.subplot(435)
plt.imshow(four);
plt.title('four')
plt.colorbar()
plt.subplot(436)
plt.imshow(five);
plt.title('five')
plt.colorbar()
plt.subplot(437)
plt.imshow(six);
plt.title('six')
plt.colorbar()
plt.subplot(438)
plt.imshow(seven);
plt.title('seven')
plt.colorbar()
plt.subplot(439)
plt.imshow(eight);
plt.title('eight')
plt.colorbar()
plt.show()

plt.imshow(nine);
plt.colorbar()
plt.title('Nine')
plt.show()

from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score
from sklearn.model_selection import train_test_split

y=df[['label']].copy()

y.head()

	label
0	0
1	1
2	2
3	3
4	4

y_test.values[0]

3

data=df.copy()
data=data.drop(['label'],axis=1)
data.head(15)

	pixel0	pixel1	pixel2	pixel3	pixel4	pixel5	pixel6	pixel7	pixel8	pixel9	pixel10	pixel11	pixel12	pixel13	pixel14	pixel15	pixel16	pixel17	pixel18	pixel19	pixel20	pixel21	pixel22	pixel23	pixel24	pixel25	pixel26	pixel27	pixel28	pixel29	pixel30	pixel31	pixel32	pixel33	pixel34	pixel35	pixel36	pixel37	pixel38	pixel39	...	pixel744	pixel745	pixel746	pixel747	pixel748	pixel749	pixel750	pixel751	pixel752	pixel753	pixel754	pixel755	pixel756	pixel757	pixel758	pixel759	pixel760	pixel761	pixel762	pixel763	pixel764	pixel765	pixel766	pixel767	pixel768	pixel769	pixel770	pixel771	pixel772	pixel773	pixel774	pixel775	pixel776	pixel777	pixel778	pixel779	pixel780	pixel781	pixel782	pixel783
0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
2	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
3	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
4	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
5	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
6	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
7	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
8	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
9	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
10	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
11	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
12	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
13	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
14	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0

15 rows × 784 columns

X_train,X_test,y_train,y_test=train_test_split(data,y,shuffle=True,test_size=0.3)

print(X_train.shape)
print(X_test.shape)

(42000, 784)
(18000, 784)

model=RandomForestClassifier(n_estimators=200,n_jobs=-1,max_features='sqrt')
model.fit(X_train,y_train)

RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',
                       max_depth=None, max_features='sqrt', max_leaf_nodes=None,
                       min_impurity_decrease=0.0, min_impurity_split=None,
                       min_samples_leaf=1, min_samples_split=2,
                       min_weight_fraction_leaf=0.0, n_estimators=100,
                       n_jobs=-1, oob_score=False, random_state=None, verbose=0,
                       warm_start=False)

y_pred=model.predict(X_test)

accuracy_score(y_test,y_pred)*100

97.87777777777778

y_test

1303     3
31753    3
1812     2
12972    2
47       7
        ..
51045    5
2349     9
41597    7
21335    5
5585     5
Name: label, Length: 18000, dtype: int64

t1=X_test.iloc[0].values
t1=t1.reshape((28,28))
actual=y_test.values[0]
predicted=y_pred[0]
plt.imshow(t1);
plt.colorbar()
plt.title('actual : '+str(actual)+" / "+'predicted: '+str(predicted))
#plt.title('actual'+'/'+'predicted')
plt.show()

sha=1211
t1=X_test.iloc[sha].values
t1=t1.reshape((28,28))
actual=y_test.values[sha]
predicted=y_pred[sha]
plt.imshow(t1);
plt.colorbar()
plt.title('actual : '+str(actual)+" / "+'predicted: '+str(predicted))
plt.show()

test.head(10)

	pixel0	pixel1	pixel2	pixel3	pixel4	pixel5	pixel6	pixel7	pixel8	pixel9	pixel10	pixel11	pixel12	pixel13	pixel14	pixel15	pixel16	pixel17	pixel18	pixel19	pixel20	pixel21	pixel22	pixel23	pixel24	pixel25	pixel26	pixel27	pixel28	pixel29	pixel30	pixel31	pixel32	pixel33	pixel34	pixel35	pixel36	pixel37	pixel38	pixel39	...	pixel744	pixel745	pixel746	pixel747	pixel748	pixel749	pixel750	pixel751	pixel752	pixel753	pixel754	pixel755	pixel756	pixel757	pixel758	pixel759	pixel760	pixel761	pixel762	pixel763	pixel764	pixel765	pixel766	pixel767	pixel768	pixel769	pixel770	pixel771	pixel772	pixel773	pixel774	pixel775	pixel776	pixel777	pixel778	pixel779	pixel780	pixel781	pixel782	pixel783
0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
2	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
3	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
4	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
5	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	255	255	255	255	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
6	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
7	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	76	22	0	0	0	0	0	0	31	60	0	0	0	0	0	0	0	0	0	0
8	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
9	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0

10 rows × 784 columns

idcode=test.id.values

test=test.drop(['id'],axis=1)

y_prediction=model.predict(test)

output=pd.DataFrame({'id':idcode,'label':y_prediction})

Now Let's Use Neural networks

path = '../content//'
path
! ls {path} -l

total 141496
-rw-r--r-- 1 root root  18401187 Sep 16 22:42 Dig-MNIST.csv
-rw-r--r-- 1 root root  13014944 Nov 25 16:33 Kannada-MNIST.zip
drwxr-xr-x 1 root root      4096 Nov 21 16:30 sample_data
-rw-r--r-- 1 root root     33899 Sep 16 22:42 sample_submission.csv
-rw-r--r-- 1 root root   8667163 Sep 16 22:42 test.csv
-rw-r--r-- 1 root root 104748945 Sep 16 22:42 train.csv

df2 = pd.read_csv('Dig-MNIST.csv')

Converting pixel to image data

class CustomImageItemList(ImageList):
    def open(self, fn):
        img = fn.reshape(28, 28)
        img = np.stack((img,)*3, axis=-1) # convert to 3 channels
        return Image(pil2tensor(img, dtype=np.float32))

    @classmethod
    def from_csv_custom(cls, path:PathOrStr, csv_name:str, imgIdx:int=1, header:str='infer', **kwargs) -> 'ItemList':
        df = pd.read_csv(Path(path)/csv_name, header=header)
        res = super().from_df(df, path=path, cols=0, **kwargs)
        # convert pixels to an ndarray
        res.items = df.iloc[:,imgIdx:].apply(lambda x: x.values / 783.0, axis=1).values
        return res

test = CustomImageItemList.from_csv_custom(path=path, csv_name='test.csv', imgIdx=1)
# DigMNIST = CustomImageItemList.from_csv_custom(path=path, csv_name='Dig-MNIST.csv')
data = (CustomImageItemList.from_csv_custom(path=path, csv_name='train.csv')
                       .split_by_rand_pct(.2)
#                       .split_by_idx(list(range(60000,70240)))
                       .label_from_df(cols='label')
                       .add_test(test, label=0)
                       .transform(get_transforms(do_flip = False))
#                        .transform(get_transforms(do_flip = False), size=49)
                       .databunch(bs=32, num_workers=16))
#                        .normalize(mnist_stats))
data

ImageDataBunch;

Train: LabelList (48000 items)
x: CustomImageItemList
Image (3, 28, 28),Image (3, 28, 28),Image (3, 28, 28),Image (3, 28, 28),Image (3, 28, 28)
y: CategoryList
1,2,3,4,5
Path: ../content;

Valid: LabelList (12000 items)
x: CustomImageItemList
Image (3, 28, 28),Image (3, 28, 28),Image (3, 28, 28),Image (3, 28, 28),Image (3, 28, 28)
y: CategoryList
4,9,1,7,8
Path: ../content;

Test: LabelList (5000 items)
x: CustomImageItemList
Image (3, 28, 28),Image (3, 28, 28),Image (3, 28, 28),Image (3, 28, 28),Image (3, 28, 28)
y: CategoryList
0,0,0,0,0
Path: ../content

data.show_batch(rows=3, figsize=(12,9))

learn = cnn_learner(data, models.resnet34, metrics=accuracy)
learn.fit_one_cycle(4)

Downloading: "https://download.pytorch.org/models/resnet34-333f7ec4.pth" to /root/.cache/torch/checkpoints/resnet34-333f7ec4.pth
100%|██████████| 83.3M/83.3M [00:00<00:00, 187MB/s]

epoch	train_loss	valid_loss	accuracy	time
0	0.605482	0.303300	0.901583	01:18
1	0.219873	0.100669	0.968750	01:18
2	0.150273	0.065184	0.981833	01:18
3	0.138628	0.059311	0.982000	01:18

learn.save('stage-1-34')  

interp = ClassificationInterpretation.from_learner(learn)

losses,idxs = interp.top_losses()

len(data.valid_ds)==len(losses)==len(idxs)

True

interp.plot_top_losses(9, figsize=(15,11))

interp.plot_confusion_matrix(figsize=(11,8), dpi=60)

interp.most_confused(min_val=2)

[(0, 1, 28),
 (7, 6, 23),
 (7, 3, 19),
 (6, 7, 12),
 (3, 5, 11),
 (8, 5, 10),
 (9, 6, 10),
 (3, 7, 9),
 (4, 5, 6),
 (5, 3, 6),
 (8, 4, 6),
 (0, 8, 5),
 (2, 3, 5),
 (5, 4, 5),
 (0, 3, 4),
 (1, 5, 4),
 (3, 4, 4),
 (6, 9, 4),
 (7, 1, 4),
 (8, 9, 4),
 (9, 4, 4),
 (2, 0, 3),
 (6, 3, 3),
 (0, 4, 2),
 (2, 5, 2),
 (2, 7, 2),
 (3, 0, 2),
 (4, 9, 2)]

learn.unfreeze()

learn.fit_one_cycle(1)

epoch	train_loss	valid_loss	accuracy	time
0	0.051757	0.016879	0.994750	01:29

learn.load('stage-1-34');

learn.lr_find()

LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.

learn.recorder.plot(suggestion=True)

Min numerical gradient: 2.29E-06
Min loss divided by 10: 2.75E-07

Here It's seen that learning rate is best between 1e-05 to 1e-04

learn.unfreeze()
learn.fit_one_cycle(2, max_lr=slice(1e-5,1e-4))

epoch	train_loss	valid_loss	accuracy	time
0	0.093132	0.030216	0.990583	01:29
1	0.056588	0.022255	0.993000	01:29

interp = ClassificationInterpretation.from_learner(learn)

losses,idxs = interp.top_losses()

len(data.valid_ds)==len(losses)==len(idxs)

True

interp.plot_top_losses(9, figsize=(15,11))

interp.plot_confusion_matrix(figsize=(10,10), dpi=60)

interp.most_confused(min_val=2)

[(0, 1, 15),
 (6, 7, 12),
 (7, 6, 7),
 (7, 3, 6),
 (6, 9, 5),
 (3, 7, 4),
 (9, 6, 4),
 (2, 0, 3),
 (3, 5, 3),
 (4, 3, 3),
 (4, 5, 3),
 (0, 4, 2),
 (2, 3, 2),
 (7, 1, 2)]

test_df = pd.read_csv(path+'test.csv')
test_df.head()

	id	pixel0	pixel1	pixel2	pixel3	pixel4	pixel5	pixel6	pixel7	pixel8	pixel9	pixel10	pixel11	pixel12	pixel13	pixel14	pixel15	pixel16	pixel17	pixel18	pixel19	pixel20	pixel21	pixel22	pixel23	pixel24	pixel25	pixel26	pixel27	pixel28	pixel29	pixel30	pixel31	pixel32	pixel33	pixel34	pixel35	pixel36	pixel37	pixel38	...	pixel744	pixel745	pixel746	pixel747	pixel748	pixel749	pixel750	pixel751	pixel752	pixel753	pixel754	pixel755	pixel756	pixel757	pixel758	pixel759	pixel760	pixel761	pixel762	pixel763	pixel764	pixel765	pixel766	pixel767	pixel768	pixel769	pixel770	pixel771	pixel772	pixel773	pixel774	pixel775	pixel776	pixel777	pixel778	pixel779	pixel780	pixel781	pixel782	pixel783
0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
2	2	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
3	3	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
4	4	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0

5 rows × 785 columns

tmp_df = pd.read_csv(path+'sample_submission.csv')
tmp_df.head()

	id	label
0	0	5
1	1	5
2	2	5
3	3	5
4	4	5

test_df.drop('id',axis = 'columns',inplace = True)

for i in range(5000):
    img = learn.data.test_ds[i][0]
    tmp_df.loc[i]=[i,int(learn.predict(img)[1])]
tmp_df

	id	label
0	0	3
1	1	0
2	2	2
3	3	6
4	4	7
...	...	...
4995	4995	1
4996	4996	0
4997	4997	1
4998	4998	6
4999	4999	3

5000 rows × 2 columns

tmp_df.to_csv('submission_resnet_34.csv',index=False)

tmp_df.head()

	id	label
0	0	3
1	1	0
2	2	2
3	3	6
4	4	7

test_df.head()

	pixel0	pixel1	pixel2	pixel3	pixel4	pixel5	pixel6	pixel7	pixel8	pixel9	pixel10	pixel11	pixel12	pixel13	pixel14	pixel15	pixel16	pixel17	pixel18	pixel19	pixel20	pixel21	pixel22	pixel23	pixel24	pixel25	pixel26	pixel27	pixel28	pixel29	pixel30	pixel31	pixel32	pixel33	pixel34	pixel35	pixel36	pixel37	pixel38	pixel39	...	pixel744	pixel745	pixel746	pixel747	pixel748	pixel749	pixel750	pixel751	pixel752	pixel753	pixel754	pixel755	pixel756	pixel757	pixel758	pixel759	pixel760	pixel761	pixel762	pixel763	pixel764	pixel765	pixel766	pixel767	pixel768	pixel769	pixel770	pixel771	pixel772	pixel773	pixel774	pixel775	pixel776	pixel777	pixel778	pixel779	pixel780	pixel781	pixel782	pixel783
0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
2	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
3	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
4	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0

5 rows × 784 columns

test1=test_df.iloc[233].values
test1=test1.reshape((28,28))
plt.imshow(test1);
plt.colorbar()
plt.show()