kernel_id
int64
24.2k
23.3M
prompt
stringlengths
8
1.85M
completetion
stringlengths
1
182k
comp_name
stringlengths
5
57
6,670,250
submission_1 = pd.read_csv('.. /input/siim-isic-melanoma-classification/sample_submission.csv') submission_1[LABELS] = predictions submission_2 = pd.read_csv('.. /input/siim-isic-melanoma-classification/sample_submission.csv') submission_2[LABELS] = predictions_2 submission = pd.read_csv('.. /input/siim-isic-melanoma...
X_test = test.drop('Survived', axis=1 ).values
Titanic - Machine Learning from Disaster
6,670,250
!pip install -q efficientnet >> /dev/null <define_variables>
log_params = dict( C = np.logspace(-5, 8, 15), penalty = ['l1', 'l2'] )
Titanic - Machine Learning from Disaster
6,670,250
DEVICE = "TPU" SEED = 42 FOLDS = 5 IMG_SIZES = [256]*FOLDS DROP_FREQ = [0.5]*FOLDS DROP_CT = [10]*FOLDS DROP_SIZE = [0.1]*FOLDS INC2019 = [0]*FOLDS INC2018 = [1]*FOLDS M1 = [1]*FOLDS M2 = [1]*FOLDS M3 = [0]*FOLDS M4 = [1]*FOLDS BATCH_SIZES = [32]*FOLDS EPOCHS = [12]*FOLDS EFF_NETS = [6]*FOLDS WGTS = [1/FOLDS]*FOLDS TTA...
log = LogisticRegression() logreg_cv = GridSearchCV(estimator=log, param_grid=log_params, cv=5) logreg_cv.fit(X, y )
Titanic - Machine Learning from Disaster
6,670,250
if DEVICE == "TPU": print("connecting to TPU...") try: tpu = tf.distribute.cluster_resolver.TPUClusterResolver() print('Running on TPU ', tpu.master()) except ValueError: print("Could not connect to TPU") tpu = None if tpu: try: print("initializing TPU...") tf.config.experimental_connect_to_cluster(tpu) tf.tpu.exp...
print("Tuned Logistic Regression Parameters: {}".format(logreg_cv.best_params_)) print("Best score is {}".format(logreg_cv.best_score_))
Titanic - Machine Learning from Disaster
6,670,250
GCS_PATH = [None]*FOLDS; GCS_PATH2 = [None]*FOLDS; GCS_PATH3 = [None]*FOLDS for i,k in enumerate(IMG_SIZES[:FOLDS]): GCS_PATH[i] = KaggleDatasets().get_gcs_path('melanoma-%ix%i'%(k,k)) GCS_PATH2[i] = KaggleDatasets().get_gcs_path('isic2019-%ix%i'%(k,k)) GCS_PATH3[i] = KaggleDatasets().get_gcs_path('malignant-v2-%ix%i'%...
log_pred = logreg_cv.predict(X_test )
Titanic - Machine Learning from Disaster
6,670,250
ROT_ = 180.0; SHR_ = 2.0 HZOOM_ = 8.0; WZOOM_ = 8.0 HSHIFT_ = 8.0; WSHIFT_ = 8.0 def get_mat(rotation, shear, height_zoom, width_zoom, height_shift, width_shift): rotation = math.pi * rotation / 180. shear = math.pi * shear / 180. def get_3x3_mat(lst): return tf.reshape(tf.concat([lst],axis=0), [3,3]) c1 = tf.math.c...
forrest_params = dict( max_depth = [n for n in range(9, 14)], min_samples_split = [n for n in range(4, 11)], min_samples_leaf = [n for n in range(2, 5)], n_estimators = [n for n in range(10, 60, 10)], )
Titanic - Machine Learning from Disaster
6,670,250
def dropout(image, DIM=256, PROBABILITY = 0.75, CT = 8, SZ = 0.2): P = tf.cast(tf.random.uniform([],0,1)<PROBABILITY, tf.int32) if(P==0)|(CT==0)|(SZ==0): return image for k in range(CT): x = tf.cast(tf.random.uniform([],0,DIM),tf.int32) y = tf.cast(tf.random.uniform([],0,DIM),tf.int32) WIDTH = tf.cast(SZ*DIM,tf.int3...
forrest = RandomForestClassifier()
Titanic - Machine Learning from Disaster
6,670,250
def read_labeled_tfrecord(example): tfrec_format = { 'image' : tf.io.FixedLenFeature([], tf.string), 'image_name' : tf.io.FixedLenFeature([], tf.string), 'patient_id' : tf.io.FixedLenFeature([], tf.int64), 'sex' : tf.io.FixedLenFeature([], tf.int64), 'age_approx' : tf.io.FixedLenFeature([], tf.int64), 'anatom_site_gene...
forest_cv = GridSearchCV(estimator=forrest, param_grid=forrest_params, cv=5) forest_cv.fit(X, y )
Titanic - Machine Learning from Disaster
6,670,250
def get_dataset(files, augment = False, shuffle = False, repeat = False, labeled=True, return_image_names=True, batch_size=16, dim=256, droprate=0, dropct=0, dropsize=0): ds = tf.data.TFRecordDataset(files, num_parallel_reads=AUTO) ds = ds.cache() if repeat: ds = ds.repeat() if shuffle: ds = ds.shuffle(1024*2) opt = ...
print("Best score: {}".format(forest_cv.best_score_)) print("Optimal params: {}".format(forest_cv.best_estimator_))
Titanic - Machine Learning from Disaster
6,670,250
def show_dataset(thumb_size, cols, rows, ds): mosaic = PIL.Image.new(mode='RGB', size=(thumb_size*cols +(cols-1), thumb_size*rows +(rows-1))) for idx, data in enumerate(iter(ds)) : img, target_or_imgid = data ix = idx % cols iy = idx // cols img = np.clip(img.numpy() * 255, 0, 255 ).astype(np.uint8) img = PIL.Image.f...
forrest_pred = forest_cv.predict(X_test) print(forrest_pred )
Titanic - Machine Learning from Disaster
6,670,250
EFNS = [efn.EfficientNetB0, efn.EfficientNetB1, efn.EfficientNetB2, efn.EfficientNetB3, efn.EfficientNetB4, efn.EfficientNetB5, efn.EfficientNetB6] def build_model(dim=128, ef=0): inp = tf.keras.layers.Input(shape=(dim,dim,3)) base = EFNS[ef](input_shape=(dim,dim,3),weights='imagenet',include_top=False) x = base(inp) ...
kaggle = pd.DataFrame({'PassengerId': passengerId, 'Survived': forrest_pred} )
Titanic - Machine Learning from Disaster
6,670,250
def get_lr_callback(batch_size=8): lr_start = 0.000005 lr_max = 0.00000125 * REPLICAS * batch_size lr_min = 0.000001 lr_ramp_ep = 5 lr_sus_ep = 0 lr_decay = 0.8 def lrfn(epoch): if epoch < lr_ramp_ep: lr =(lr_max - lr_start)/ lr_ramp_ep * epoch + lr_start elif epoch < lr_ramp_ep + lr_sus_ep: lr = lr_max else: lr =(lr_m...
filename = 'submit.csv' kaggle.to_csv(filename, index=False )
Titanic - Machine Learning from Disaster
847,197
%.2f'%y,size=14) \ <define_variables>
train=pd.read_csv('.. /input/train.csv') test=pd.read_csv('.. /input/test.csv' )
Titanic - Machine Learning from Disaster
847,197
VERBOSE = 0 DISPLAY_PLOT = True oof_pred = []; oof_tar = []; oof_val = []; oof_names = []; oof_folds = [] preds = np.zeros(( count_data_items(files_test),1)) skf = KFold(n_splits=FOLDS,shuffle=True,random_state=SEED) oof_pred = []; oof_tar = []; oof_val = []; oof_names = []; oof_folds = [] preds = np.zeros(( count_dat...
train.drop(labels='Cabin',inplace=True,axis=1) test.drop(labels='Cabin',inplace=True,axis=1)
Titanic - Machine Learning from Disaster
847,197
oof = np.concatenate(oof_pred); true = np.concatenate(oof_tar); names = np.concatenate(oof_names); folds = np.concatenate(oof_folds) auc = roc_auc_score(true,oof) print('Overall OOF AUC with TTA = %.3f'%auc) df_oof = pd.DataFrame(dict( image_name = names, target=true, pred = oof, fold=folds)) df_oof.to_csv('oof.csv...
def check_class(x): if pd.isnull(x['Age']): return pmean[x['Pclass']] return x['Age'] pmean=train.groupby('Pclass' ).mean() ['Age'] train['Age']=train.apply(check_class,axis=1) test['Age']=test.apply(check_class,axis=1)
Titanic - Machine Learning from Disaster
847,197
if INFER_TEST: ds = get_dataset(files_test, augment=False, repeat=False, dim=IMG_SIZES[fold], labeled=False, return_image_names=True) image_names = np.array([img_name.numpy().decode("utf-8") for img, img_name in iter(ds.unbatch())]) submission = pd.DataFrame(dict(image_name=image_names, target=preds[:,0])) submissio...
test['Fare']=test['Fare'].fillna(np.mean(test['Fare'])).astype(float)
Titanic - Machine Learning from Disaster
847,197
train = pd.read_csv('/kaggle/input/siim-isic-melanoma-classification/train.csv') test = pd.read_csv('/kaggle/input/siim-isic-melanoma-classification/test.csv') sub = pd.read_csv('/kaggle/input/siim-isic-melanoma-classification/sample_submission.csv') <define_variables>
vtrain=train vtest=test
Titanic - Machine Learning from Disaster
847,197
models = [ "getting-started-with-tfrecords", "melanoma-efficientnetb6-with-attention-mechanism", "triple-stratified-kfold-with-tfrecords"] models = ["384-E6-with-2018","512-E6","768-E2","512-E5","effb3-fulldata-upsample","effb2-fulldata-upsample"] models = ["effb0-fulldata-upsample","effb1-fulldata-upsample","effb2-512...
train=train.loc[:,['Pclass','Survived','Sex','Age','SibSp','Parch','Fare','Embarked']] test=test.loc[:,['Pclass','Sex','Age','SibSp','Parch','Fare','Embarked']]
Titanic - Machine Learning from Disaster
847,197
train["pred_rank"] = 0 train["pred_power"] = 0 train["pred_avg"] = 0 for c in models: train["pred_rank"] += train[c].rank() / train[c].rank().max() train["pred_power"] += np.power(train[c],2)/np.power(train[c],2 ).max() train["pred_avg"] += train [c]/train [c].max() train["pred_rank"] = train["pred_rank"]/len(models) ...
pc=pd.get_dummies(train['Pclass'],drop_first=True,prefix='pclass') pctest=pd.get_dummies(test['Pclass'],drop_first=True,prefix='pclass') sex=pd.get_dummies(train['Sex'],drop_first=True,prefix='sex') sextest=pd.get_dummies(test['Sex'],drop_first=True,prefix='sex') em=pd.get_dummies(train['Embarked'],drop_first=True)...
Titanic - Machine Learning from Disaster
847,197
test["target"] = 0.0 for c in models: test["target"] += test[c].rank() / test[c].rank().max() test["target"] = test["target"]/len(models) sub = test[["image_name","target"]] sub.to_csv("submission_rank.csv",index=False) sub.head()<save_to_csv>
def fill(x): for i in range(len(unique_surnames_train)) : if unique_surnames_train[i] in x: return i extratrain=pd.get_dummies(vtrain["Name"].apply(fill ).replace(np.arange(len(unique_surnames_train)) ,unique_surnames_train),drop_first=True) extratest=pd.get_dummies(vtest["Name"].apply(fill ).replace(np.arange(len(uni...
Titanic - Machine Learning from Disaster
847,197
test["target"] = 0.0 for c in models: test["target"] += np.power(test[c],2)/np.power(test[c],2 ).max() test["target"] = test["target"]/len(models) sub = test[["image_name","target"]] sub.to_csv("submission_pow.csv",index=False) sub.head()<save_to_csv>
temp=pd.concat([extratest,pd.DataFrame(np.zeros(( test.shape[0],9)) ,columns=['Col.','Don.','Jonkheer.','Lady.','Major.','Mlle.','Mme.','Sir.', 'the'] ).astype('int')],axis=1)
Titanic - Machine Learning from Disaster
847,197
test["target"] = 0.0 for c in models: test["target"] += test[c]/test[c].max() test["target"] = test["target"]/len(models) sub = test[["image_name","target"]] sub.to_csv("submission_avg.csv",index=False) sub.head()<compute_train_metric>
def hasalpha(x): for i in x: if str.isalpha(i): return i return 'non' extr=pd.get_dummies(vtrain['Ticket'].apply(hasalpha)).iloc[:,:-1] exte=pd.get_dummies(vtest['Ticket'].apply(hasalpha)).iloc[:,:-1].loc[:,extr.columns]
Titanic - Machine Learning from Disaster
847,197
def dim_optimizer(df_oof, features, init_points = 20, n_iter = 30): pbounds = {'c0':(0.0, 1.0), 'c1':(0.0, 1.0), 'c2':(0.0, 1.0),'c3':(0.0, 1.0),'c4':(0.0, 1.0),'c5':(0.0, 1.0)} features = features def dim_opt(df_oof, c0,c1,c2,c3,c4,c5): x = c0*df_oof[ features[0] ] + c1*df_oof[ features[1]] + c2*df_oof[ features[2]] +...
train=pd.concat([train,pc,sex,em],axis=1 ).drop(['Pclass','Sex','Embarked'],axis=1) test=pd.concat([test,pctest,sextest,emtest],axis=1 ).drop(['Pclass','Sex','Embarked'],axis=1 )
Titanic - Machine Learning from Disaster
847,197
def bo_pred(df): x = c0*df[ models[0] ] + c1*df[ models[1]] + c2*df[ models[2]] + c3*df[ models[3]] + c4*df[ models[4]] + c5*df[ models[5]] return x train["pred"] = bo_pred(train) score = metrics.roc_auc_score(train['target'], train['pred']) print(f"auc bo:{score}") <save_to_csv>
dftr=pd.read_csv('.. /input/train.csv') dfte=pd.read_csv('.. /input/test.csv') def app(x): if pd.notnull(x): return x[0] return xx=pd.get_dummies(dftr['Cabin'].apply(app)).iloc[:,:-2] xte=pd.get_dummies(dfte['Cabin'].apply(app)) xte['T']=np.zeros(( len(xte),1)).astype('int') xte=xte.iloc[:,:-2]
Titanic - Machine Learning from Disaster
847,197
test["target"] = bo_pred(test) sub = test[["image_name","target"]] sub.to_csv("submission_bo.csv",index=False) sub.head()<import_modules>
train=pd.concat([train,extratrain,extr,xx],axis=1) test=pd.concat([test,extratest,exte,xte],axis=1)
Titanic - Machine Learning from Disaster
847,197
!pip install -q efficientnet <load_from_csv>
xtrain=train.iloc[:,1:].values xtest=test.values ytrain=train.iloc[:,0].values
Titanic - Machine Learning from Disaster
847,197
train = pd.read_csv('/kaggle/input/siim-isic-melanoma-classification/train.csv') test = pd.read_csv('/kaggle/input/siim-isic-melanoma-classification/test.csv') sample = pd.read_csv('/kaggle/input/siim-isic-melanoma-classification/sample_submission.csv') train.head()<set_options>
sc_x=StandardScaler() xtrain=sc_x.fit_transform(xtrain) xtest=sc_x.transform(xtest)
Titanic - Machine Learning from Disaster
847,197
AUTO = tf.data.experimental.AUTOTUNE try: tpu = tf.distribute.cluster_resolver.TPUClusterResolver() print('Running on TPU ', tpu.master()) except ValueError: tpu = None if tpu: tf.config.experimental_connect_to_cluster(tpu) tf.tpu.experimental.initialize_tpu_system(tpu) strategy = tf.distribute.experimental.TPUStrat...
regressor=LogisticRegression(C=1,solver='saga') regressor.fit(xtrain,ytrain )
Titanic - Machine Learning from Disaster
847,197
SEED = 42 FOLDS=3 EFF_NETS = [6]*FOLDS BATCH_SIZES = [bs * strategy.num_replicas_in_sync for bs in [32]*FOLDS] IMG_SIZES = [512]*FOLDS EPOCHS = [10]*FOLDS DROPOUT = 0.25 LR = 0.00004 WARMUP = 5 CLASS_WEIGHT = {0: train['benign_malignant'].value_counts().malignant/len(train), 1: train['benign_malignant'].value_counts()....
regressor2=SVC(C=1,gamma=0.01,kernel='rbf') regressor2.fit(xtrain,ytrain )
Titanic - Machine Learning from Disaster
847,197
DATASET = {512: '512x512-melanoma-tfrecords-70k-images', 384: 'melanoma-384x384', 192: 'melanoma-192x192'} GCS_PATH = [None]*FOLDS; GCS_PATH2 = [None]*FOLDS for i,k in enumerate(IMG_SIZES): GCS_PATH[i] = KaggleDatasets().get_gcs_path(DATASET[IMG_SIZES[0]]) GCS_PATH2[i] = KaggleDatasets().get_gcs_path('isic2019-%ix%i'%...
regressor3=RandomForestClassifier(criterion='entropy',n_estimators=500) regressor3.fit(xtrain,ytrain )
Titanic - Machine Learning from Disaster
847,197
def lrfn(epoch): if epoch < lr_ramp_ep: lr =(lr_max - lr_start)/ lr_ramp_ep * epoch + lr_start elif epoch < lr_ramp_ep + lr_sus_ep: lr = lr_max else: lr =(lr_max - lr_min)* lr_decay**(epoch - lr_ramp_ep - lr_sus_ep)+ lr_min return lr def get_lr_callback(batch_size=8): lr_start = 0.000005 lr_max = 0.000003 * batch_size ...
regressor3.score(xtrain,ytrain )
Titanic - Machine Learning from Disaster
847,197
lr_start = 0.000005 lr_max = 0.000003 * BATCH_SIZES[0] lr_min = 0.000001 lr_ramp_ep = 5 lr_sus_ep = 0 lr_decay = 0.3 def lrfn(epoch): if epoch < lr_ramp_ep: lr =(lr_max - lr_start)/ lr_ramp_ep * epoch + lr_start elif epoch < lr_ramp_ep + lr_sus_ep: lr = lr_max else: lr =(lr_max - lr_min)* lr_decay**(epoch - lr_ramp_ep ...
test['Survived']=regressor.predict(xtest )
Titanic - Machine Learning from Disaster
847,197
ROT_ = 180.0 SHR_ = 2.0 HZOOM_ = 8.0 WZOOM_ = 8.0 HSHIFT_ = 8.0 WSHIFT_ = 8.0 def get_mat(rotation, shear, height_zoom, width_zoom, height_shift, width_shift): rotation = math.pi * rotation / 180. shear = math.pi * shear / 180. def get_3x3_mat(lst): return tf.reshape(tf.concat([lst],axis=0), [3,3]) c1 = tf.math.cos(...
param_svm=[ { 'kernel':['rbf'], 'C':[0.1,0.01,1,5,10], 'gamma':[1,0.1,0.01] }, { 'kernel':['linear'], 'C':[0.1,0.01,1,5,10], 'gamma':[1,0.1,0.01,5,10] }, { 'kernel':['sigmoid'], 'C':[0.1,0.01,1,5,10], 'gamma':[1,0.1,0.01] } ] param_rf=[ { 'n_estimators':[10,100,300,600,500], 'criterion':['gini'],'max_depth':[2,5,10,20,...
Titanic - Machine Learning from Disaster
847,197
def decode_image(image): image = tf.image.decode_jpeg(image, channels=3) image = tf.cast(image, tf.float32)/255.0 image = tf.reshape(image, [*IMG_SIZES[0:2], 3]) return image def read_tfrecord(example, labeled, return_imgname=False): tfrecord_format = { "image": tf.io.FixedLenFeature([], tf.string), "target": tf.io.F...
print("SVM") print(gc_svm.best_params_) print('-----------------------------------------------------------------------------------') print('logistic regression') print(gc_logistic.best_params_) print('-----------------------------------------------------------------------------------') print('random forest') pri...
Titanic - Machine Learning from Disaster
847,197
def show_dataset(thumb_size, cols, rows, ds): mosaic = PIL.Image.new(mode='RGB', size=(thumb_size*cols +(cols-1), thumb_size*rows +(rows-1))) for idx, data in enumerate(iter(ds)) : img, target_or_imgid = data ix = idx % cols iy = idx // cols img = np.clip(img.numpy() * 255, 0, 255 ).astype(np.uint8) img = PIL.Image.f...
print("SVM") print(gc_svm.best_score_) print('-----------------------------------------------------------------------------------') print('logistic regression') print(gc_logistic.best_score_) print('-----------------------------------------------------------------------------------') print('random forest') print...
Titanic - Machine Learning from Disaster
847,197
plot_transform(7 )<define_variables>
output_train=pd.DataFrame([regressor.predict(xtrain),regressor2.predict(xtrain)] ).apply(lambda x: x.mode() ).iloc[0].values output=pd.DataFrame([regressor.predict(xtest),regressor2.predict(xtest)] ).apply(lambda x: x.mode() )
Titanic - Machine Learning from Disaster
847,197
VERBOSE = 2 DISPLAY_PLOT = True skf = KFold(n_splits=FOLDS,shuffle=True,random_state=SEED) oof_pred = []; oof_tar = []; oof_val = []; oof_names = []; oof_folds = [] preds = np.zeros(( count_data_items(test_filenames),1)) for fold,(idxT,idxV)in enumerate(skf.split(np.arange(15))): if tpu: tf.tpu.experimental.initialize...
print(classification_report(ytrain,output_train))
Titanic - Machine Learning from Disaster
847,197
oof = np.concatenate(oof_pred); true = np.concatenate(oof_tar); names = np.concatenate(oof_names); folds = np.concatenate(oof_folds) auc = roc_auc_score(true,oof) print('Overall OOF AUC with TTA = %.3f'%auc) df_oof = pd.DataFrame(dict( image_name = names, target=true, pred = oof, fold=folds)) df_oof.to_csv('oof.csv...
submission=pd.read_csv('.. /input/gender_submission.csv') submission['Survived']=output.iloc[0].values
Titanic - Machine Learning from Disaster
847,197
<save_to_csv><EOS>
submission.to_csv('finaloutput.csv',index=False )
Titanic - Machine Learning from Disaster
10,022,452
<SOS> metric: categorizationaccuracy Kaggle data source: titanic-machine-learning-from-disaster<load_from_csv>
%matplotlib inline for dirname, _, filenames in os.walk('/kaggle/input'): for filename in filenames: print(os.path.join(dirname, filename))
Titanic - Machine Learning from Disaster
10,022,452
train = pd.read_csv('/kaggle/input/melanomaextendedtabular/external_upsampled_tabular.csv') train.head()<categorify>
sns.set(style="darkgrid") train_full = pd.read_csv('.. /input/titanic/train.csv', index_col='PassengerId') test_full = pd.read_csv('.. /input/titanic/test.csv', index_col='PassengerId') train_NA = train_full.isna().sum() test_NA = test_full.isna().sum() pd.concat([train_NA, test_NA], axis=1, sort = False, keys=['Tra...
Titanic - Machine Learning from Disaster
10,022,452
train['anatom_site_general_challenge'].replace(['anterior torso','lateral torso','posterior torso'], 'torso', inplace=True )<count_missing_values>
null_fare = test_full[test_full.Fare.isnull() ].index[0] test_full.loc[null_fare, 'Fare'] = 0
Titanic - Machine Learning from Disaster
10,022,452
train.isnull().sum()<count_missing_values>
y = train_full.Survived train_X_full = train_full[train_full.columns.drop('Survived')] combined = pd.concat([train_full, test_full], axis=0 )
Titanic - Machine Learning from Disaster
10,022,452
train.isnull().sum()<data_type_conversions>
def alone(df): if('SibSp' in df.columns)and('Parch' in df.columns): df['Family'] = df['SibSp'] + df['Parch'] + 1 le = LabelEncoder() df['le_Ticket'] = le.fit_transform(df['Ticket']) df['Same_Ticket'] = df.duplicated(['le_Ticket']) df['Alone'] = np.where(( df['Family'] > 1)|(df['Same_Ticket']), False, True) plt.figur...
Titanic - Machine Learning from Disaster
10,022,452
def fillna(df, column): na_idx = df[df[column].isnull() ].index.tolist() prob = df[column].value_counts(normalize=True ).sort_index().tolist() for i in na_idx: df.iloc[i, df.columns.get_loc(column)] = choices(sorted(df[column].dropna().unique().tolist()), prob)[0] train.sex.replace('unknown', np.nan, inplace=True) fil...
combined = alone(combined) train_X_full = combined[combined.Survived.notnull() ] test_full = combined[combined.Survived.isnull() ]
Titanic - Machine Learning from Disaster
10,022,452
train = pd.get_dummies(train, columns=["anatom_site_general_challenge"], prefix='site', drop_first=True) train.replace({'sex': {'female':0, 'male': 1}}, inplace=True) train.head()<categorify>
def fix_cabin(df): t = df.Cabin.fillna('U') df['Cabin'] = t.str.slice(0,1) plt.figure(figsize=(20, 15)) sns.catplot(x="Cabin", kind="count", palette="ch:.25", data=df) plt.title('Number of passengers per cabin')
Titanic - Machine Learning from Disaster
10,022,452
test = pd.get_dummies(test, columns=["anatom_site_general_challenge"], prefix='site', drop_first=True) test.replace({'sex': {'female':0, 'male': 1}}, inplace=True) test.head()<normalization>
fix_cabin(combined) train_X_full = combined[combined.Survived.notnull() ] test_full = combined[combined.Survived.isnull() ]
Titanic - Machine Learning from Disaster
10,022,452
y_train = train['target'] X_train = train.drop(['image_name','height','width','target'], axis=1) X_test = test.drop(['image_name','patient_id'],axis=1) scaler = StandardScaler() X_train_s = X_train.copy() X_train_s[['age_approx']] = scaler.fit_transform(X_train_s[['age_approx']]) X_test_s = X_test.copy() X_test_s[['...
fix_embark(combined) train_X_full = combined[combined.Survived.notnull() ] test_full = combined[combined.Survived.isnull() ]
Titanic - Machine Learning from Disaster
10,022,452
lr = LogisticRegression(penalty='l2', class_weight='balanced', random_state=SEED) nb = GaussianNB() rf = RandomForestClassifier(n_estimators=1000, max_depth=2, class_weight='balanced', n_jobs=-1, random_state=SEED) estimators = [lr, rf, nb] cv = StratifiedKFold(5, shuffle=True, random_state=SEED) def model_cv(X_trai...
Titanic - Machine Learning from Disaster
10,022,452
def model_blend(X_train, y_train, X_test, estimators): mean_prob = 0 for est in estimators: est.fit(X_train, y_train) mean_prob += est.predict_proba(X_test)[:,1] return mean_prob/len(estimators) meta_df = pd.DataFrame(columns=['image_name', 'target']) meta_df['image_name'] = sample['image_name'] meta_df['target'] = ...
def extract_title(df): if 'Name' in df.columns: df['Title'] = df['Name'].str.split(',', expand=True)[1].str.split('.', expand=True)[0].str.strip() df = df[df.columns.drop('Name')] df['Title'] = df['Title'].replace({'Ms': 'Miss', 'Mlle': 'Miss', 'Mme': 'Mrs', 'Lady': 'Mrs', 'the Countess': 'Mrs', 'Dona': 'Mrs', 'Don': '...
Titanic - Machine Learning from Disaster
10,022,452
submission_effnet_ensemble = pd.read_csv('.. /input/effnet-ensemble/submission_effnet_ensemble.csv') submission.target =(submission_effnet_ensemble.target)+(meta_df.target * 0.1) submission.to_csv('submission.csv', index=False) submission.head()<import_modules>
temp = train_X_full.copy() temp['Survived'] = y print(temp['Title'].unique())
Titanic - Machine Learning from Disaster
10,022,452
import numpy as np import pandas as pd import numpy as np import pandas as pd import os <load_from_csv>
def extract_ticket(df): df['Ticket_Letters'] = df['Ticket'].str.replace('\d+', '') df.loc[df['Ticket_Letters']=='','Ticket_Letters'] = 'NA' df.drop(columns=['Ticket'], inplace=True) return df
Titanic - Machine Learning from Disaster
10,022,452
test = pd.read_csv('.. /input/siim-isic-melanoma-classification/test.csv') first = pd.read_csv('.. /input/output-of-best-public-submission/submission_best.csv') second = pd.read_csv('.. /input/output-of-best-public-submission/submission_first.csv') third = pd.read_csv('.. /input/output-of-best-public-submission/subm...
combined = extract_ticket(combined) train_X_full = combined[combined.Survived.notnull() ] test_full = combined[combined.Survived.isnull() ] train_X_full.head()
Titanic - Machine Learning from Disaster
10,022,452
arg1 =(2/3)*first['target'] arg2 =(1/6)*second['target'] arg3 =(1/6)*third['target'] submission['target'] = arg1 + arg2 + arg3 submission.to_csv('submission.csv', index=False) submission.head()<install_modules>
train_X_full = combined[combined.Survived.notnull() ] test_full = combined[combined.Survived.isnull() ]
Titanic - Machine Learning from Disaster
10,022,452
!pip install -q efficientnet >> /dev/null<import_modules>
temp = combined.copy() to_one_hot = ['Sex', 'Embarked', 'Ticket_Letters', 'Cabin', 'Title', 'Alone', 'FareBin'] temp = pd.concat([temp, pd.get_dummies(temp[to_one_hot])], axis=1) temp.drop(columns=to_one_hot, axis=1, inplace=True) combined = temp combined.head()
Titanic - Machine Learning from Disaster
10,022,452
import json import pandas as pd, numpy as np from kaggle_datasets import KaggleDatasets import tensorflow as tf, re, math import tensorflow.keras.backend as K import efficientnet.tfkeras as efn from sklearn.model_selection import KFold from sklearn.metrics import roc_auc_score import matplotlib.pyplot as plt<define_sea...
train_X_full = combined[combined.Survived.notnull() ] test_full = combined[combined.Survived.isnull() ]
Titanic - Machine Learning from Disaster
10,022,452
DEVICE = "TPU" SEED = 27 FOLDS = 5 IMG_SIZES = [512,512,512,512,512] INC2019 = [1,1,1,1,1] INC2018 = [1,1,1,1,1] BATCH_SIZES = [32]*FOLDS EPOCHS = [20]*FOLDS EFF_NETS = [6,6,6,6,6] WGTS = [1/FOLDS]*FOLDS TTA = 11<choose_model_class>
def convert_cat(df): le = LabelEncoder() le_train_X = df.copy() s = df.dtypes=='object' cat_features = list(s[s].index) for col in cat_features: le_train_X[col] = le.fit_transform(df[col]) return le_train_X
Titanic - Machine Learning from Disaster
10,022,452
if DEVICE == "TPU": print("connecting to TPU...") try: tpu = tf.distribute.cluster_resolver.TPUClusterResolver() print('Running on TPU ', tpu.master()) except ValueError: print("Could not connect to TPU") tpu = None if tpu: try: print("initializing TPU...") tf.config.experimental_connect_to_cluster(tpu) tf.tpu.exp...
without_survived = combined.copy() without_survived.drop(columns=['Survived'], inplace=True) without_survived.head()
Titanic - Machine Learning from Disaster
10,022,452
GCS_PATH = [None]*FOLDS; GCS_PATH2 = [None]*FOLDS for i,k in enumerate(IMG_SIZES): GCS_PATH[i] = KaggleDatasets().get_gcs_path('melanoma-%ix%i'%(k,k)) GCS_PATH2[i] = KaggleDatasets().get_gcs_path('isic2019-%ix%i'%(k,k)) print(GCS_PATH) print() print(GCS_PATH2) files_train = np.sort(np.array(tf.io.gfile.glob(GCS_PATH[...
missing_age_model = predict_age(without_survived )
Titanic - Machine Learning from Disaster
10,022,452
ROT_ = 180.0 SHR_ = 2.0 HZOOM_ = 8.0 WZOOM_ = 8.0 HSHIFT_ = 8.0 WSHIFT_ = 8.0<normalization>
missing_age = without_survived[without_survived.Age.isnull() ] missing_age = missing_age[missing_age.columns.drop('Age')] pred_age = missing_age_model.predict(missing_age) output_age = pd.DataFrame({'Age': pred_age}, index=missing_age.index) output_age.transpose()
Titanic - Machine Learning from Disaster
10,022,452
def get_mat(rotation, shear, height_zoom, width_zoom, height_shift, width_shift): rotation = math.pi * rotation / 180. shear = math.pi * shear / 180. def get_3x3_mat(lst): return tf.reshape(tf.concat([lst],axis=0), [3,3]) c1 = tf.math.cos(rotation) s1 = tf.math.sin(rotation) one = tf.constant([1],dtype='float32') ...
test = combined.copy() test = test.combine_first(output_age) pd.DataFrame(test.Age ).transpose()
Titanic - Machine Learning from Disaster
10,022,452
def read_labeled_tfrecord(example): tfrec_format = { 'image' : tf.io.FixedLenFeature([], tf.string), 'image_name' : tf.io.FixedLenFeature([], tf.string), 'patient_id' : tf.io.FixedLenFeature([], tf.int64), 'sex' : tf.io.FixedLenFeature([], tf.int64), 'age_approx' : tf.io.FixedLenFeature([], tf.int64), 'anatom_site_gene...
combined = test train_X_full = combined[combined.Survived.notnull() ] test_full = combined[combined.Survived.isnull() ]
Titanic - Machine Learning from Disaster
10,022,452
def get_dataset(files, augment = False, shuffle = False, repeat = False, labeled=True, return_image_names=True, batch_size=16, dim=256): ds = tf.data.TFRecordDataset(files, num_parallel_reads=AUTO) ds = ds.cache() if repeat: ds = ds.repeat() if shuffle: ds = ds.shuffle(1024*8) opt = tf.data.Options() opt.experimental...
y = train_X_full.Survived train_X_full = train_X_full.drop(columns=['Survived'], axis=1) train_X, valid_X, train_y, valid_y = train_test_split(train_X_full, y, random_state=5 )
Titanic - Machine Learning from Disaster
10,022,452
EFNS = [efn.EfficientNetB0, efn.EfficientNetB1, efn.EfficientNetB2, efn.EfficientNetB3, efn.EfficientNetB4, efn.EfficientNetB5, efn.EfficientNetB6] def build_model(dim=128, ef=6): inp = tf.keras.layers.Input(shape=(dim,dim,3)) base = EFNS[ef](input_shape=(dim,dim,3),weights='imagenet',include_top=False) x = base(inp) ...
def scores(results): key_min = min(results.keys() , key=(lambda k: results[k])) key_max = max(results.keys() , key=(lambda k: results[k])) print('Highest score at %d of %.4f' %(key_max, results[key_max])) print('Lowest score at %d of %.4f' %(key_min, results[key_min])) return key_max, key_min
Titanic - Machine Learning from Disaster
10,022,452
def get_lr_callback(batch_size=8): lr_start = 0.000005 lr_max = 0.00000125 * REPLICAS * batch_size lr_min = 0.000001 lr_ramp_ep = 5 lr_sus_ep = 0 lr_decay = 0.8 def lrfn(epoch): if epoch < lr_ramp_ep: lr =(lr_max - lr_start)/ lr_ramp_ep * epoch + lr_start elif epoch < lr_ramp_ep + lr_sus_ep: lr = lr_max else: lr =(lr_m...
print('Number of trees: ') high, low = scores(results )
Titanic - Machine Learning from Disaster
10,022,452
fold = 3 VERBOSE = 2 DISPLAY_PLOT = True skf = KFold(n_splits=FOLDS,shuffle=True,random_state=SEED) oof_pred = []; oof_tar = []; oof_val = []; oof_names = []; oof_folds = [] preds = np.zeros(( count_data_items(files_test),1)) if DEVICE=='TPU': if tpu: tf.tpu.experimental.initialize_tpu_system(tpu) print(' print(' (I...
best_rf_model = RandomForestClassifier(n_estimators=high,random_state=0 ).fit(train_X, train_y) pred_valid1 = best_rf_model.predict(valid_X) print('Mean absolute error: \t%.4f' %mean_absolute_error(pred_valid1, valid_y)) print('Accuracy score: \t%.4f' %accuracy_score(valid_y, pred_valid1))
Titanic - Machine Learning from Disaster
10,022,452
ds = get_dataset(files_test, augment=False, repeat=False, dim=IMG_SIZES[fold], labeled=False, return_image_names=True) image_names = np.array([img_name.numpy().decode("utf-8") for img, img_name in iter(ds.unbatch())]) submission = pd.DataFrame(dict(image_name=image_names, target=preds[:,0])) submission = submission....
perm = PermutationImportance(best_rf_model, random_state=1 ).fit(valid_X, valid_y) eli5.show_weights(perm, feature_names=valid_X.columns.tolist() )
Titanic - Machine Learning from Disaster
10,022,452
warnings.filterwarnings('ignore' )<load_from_csv>
results1 = {} for i in range(1, len(train_X.columns)) : selector = SelectKBest(f_classif, k=i) X_new = selector.fit_transform(train_X[feature_cols], train_y) selected_features = pd.DataFrame(selector.inverse_transform(X_new), index=train_X.index, columns=feature_cols) selected_cols = selected_features.columns[select...
Titanic - Machine Learning from Disaster
10,022,452
train = pd.read_csv("/kaggle/input/siim-isic-melanoma-classification/train.csv") test = pd.read_csv("/kaggle/input/siim-isic-melanoma-classification/test.csv") sample = pd.read_csv("/kaggle/input/siim-isic-melanoma-classification/sample_submission.csv" )<install_modules>
high1, low1 = scores(results1 )
Titanic - Machine Learning from Disaster
10,022,452
!pip install -q efficientnet<import_modules>
results2 = {} for i in range(1, len(train_X.columns)) : selector = SelectKBest(f_classif, k=i) X_new = selector.fit_transform(train_X[feature_cols], train_y) selected_features = pd.DataFrame(selector.inverse_transform(X_new), index=train_X.index, columns=feature_cols) selected_cols = selected_features.columns[select...
Titanic - Machine Learning from Disaster
10,022,452
import tensorflow as tf import tensorflow.keras.backend as K import efficientnet.tfkeras as efn from kaggle_datasets import KaggleDatasets<define_variables>
high2, low2 = scores(results2 )
Titanic - Machine Learning from Disaster
10,022,452
GCS_PATH = KaggleDatasets().get_gcs_path('melanoma-512x512') GCS_PATH2 = KaggleDatasets().get_gcs_path('malignant-v2-512x512') GCS_PATH3 = KaggleDatasets().get_gcs_path('isic2019-512x512') filenames_train1 = tf.io.gfile.glob(GCS_PATH + '/train*.tfrec') filenames_train2 = tf.io.gfile.glob(GCS_PATH2 + '/train%.2i*.tf...
results3 = {} for i in range(1, len(train_X.columns)) : selector = SelectKBest(f_classif, k=i) X_new = selector.fit_transform(train_X[feature_cols], train_y) selected_features = pd.DataFrame(selector.inverse_transform(X_new), index=train_X.index, columns=feature_cols) selected_cols = selected_features.columns[select...
Titanic - Machine Learning from Disaster
10,022,452
filenames_train = np.array(filenames_train1+filenames_train2+filenames_train3) np.random.shuffle(filenames_train) np.random.shuffle(filenames_train )<feature_engineering>
high3, low3 = scores(results3 )
Titanic - Machine Learning from Disaster
10,022,452
AUTO = tf.data.experimental.AUTOTUNE<set_options>
xg_model = xgb.XGBClassifier() results4 = {} for i in range(1, len(train_X.columns)) : selector = SelectKBest(f_classif, k=i) X_new = selector.fit_transform(train_X[feature_cols], train_y) selected_features = pd.DataFrame(selector.inverse_transform(X_new), index=train_X.index, columns=feature_cols) selected_cols = s...
Titanic - Machine Learning from Disaster
10,022,452
cfg = dict( batch_size=32, img_size=512, lr_start=0.000005, lr_max=0.00000125, lr_min=0.000001, lr_rampup=5, lr_sustain=0, lr_decay=0.8, epochs=12, transform_prob=1.0, rot=180.0, shr=2.0, hzoom=8.0, wzoom=8.0, hshift=8.0, wshift=8.0, optimizer='adam', label_smooth_fac=0.05, tta_steps=20 )<normalization>
high4, low4 = scores(results4 )
Titanic - Machine Learning from Disaster
10,022,452
def get_mat(rotation, shear, height_zoom, width_zoom, height_shift, width_shift): rotation = math.pi * rotation / 180. shear = math.pi * shear / 180. c1 = tf.math.cos(rotation) s1 = tf.math.sin(rotation) one = tf.constant([1], dtype='float32') zero = tf.constant([0], dtype='float32') rotation_matrix = tf.reshape(...
feature_cols = train_X.columns selector = SelectKBest(f_classif, k=high1) X_new = selector.fit_transform(train_X[feature_cols], train_y) selected_features = pd.DataFrame(selector.inverse_transform(X_new), index=train_X.index, columns=feature_cols) selected_cols = selected_features.columns[selected_features.var() != ...
Titanic - Machine Learning from Disaster
10,022,452
def transform(image, cfg): DIM = cfg['img_size'] XDIM = DIM % 2 rot = cfg['rot'] * tf.random.normal([1], dtype='float32') shr = cfg['shr'] * tf.random.normal([1], dtype='float32') h_zoom = 1.0 + tf.random.normal([1], dtype='float32')/ cfg['hzoom'] w_zoom = 1.0 + tf.random.normal([1], dtype='float32')/ cfg['wzoom'] h_...
feature_cols = train_X.columns selector = SelectKBest(f_classif, k=high4) X_new = selector.fit_transform(train_X[feature_cols], train_y) selected_features = pd.DataFrame(selector.inverse_transform(X_new), index=train_X.index, columns=feature_cols) selected_cols2 = selected_features.columns[selected_features.var() !=...
Titanic - Machine Learning from Disaster
10,022,452
def dropout(image, DIM=512, PROBABILITY = 0.75, CT = 8, SZ = 0.2): P = tf.cast(tf.random.uniform([],0,1)<PROBABILITY, tf.int32) if(P==0)|(CT==0)|(SZ==0): return image for k in range(CT): x = tf.cast(tf.random.uniform([],0,DIM),tf.int32) y = tf.cast(tf.random.uniform([],0,DIM),tf.int32) WIDTH = tf.cast(SZ*DIM,tf.int3...
cv_results = {} for i in range(2, 10): cv_score = cross_val_score(xg_best_model, cv_X, cv_y, cv=i) cv_results[i] = cv_score.mean()
Titanic - Machine Learning from Disaster
10,022,452
def prepare_image(img, cfg=None,droprate=0.5,dropct=8,dropsize=0.2): img = tf.image.decode_jpeg(img, channels=3) img = tf.image.resize(img, [cfg['img_size'], cfg['img_size']], antialias=True) img = tf.cast(img, tf.float32)/ 255.0 if cfg['transform_prob'] > tf.random.uniform([1], minval=0, maxval=1): img = transform(i...
cv_high, cv_low = scores(cv_results )
Titanic - Machine Learning from Disaster
10,022,452
def read_labeled_tfrecord(example): LABELED_TFREC_FORMAT = { 'image': tf.io.FixedLenFeature([], tf.string), 'image_name': tf.io.FixedLenFeature([], tf.string), 'target': tf.io.FixedLenFeature([], tf.int64) } example = tf.io.parse_single_example(example, LABELED_TFREC_FORMAT) return example['image'], example['target']...
cv_score = cross_val_score(best_k_model, cv_X, cv_y, cv=cv_high) print('Mean cross-validation score: %.2f' %(cv_score.mean() *100))
Titanic - Machine Learning from Disaster
10,022,452
def read_unlabeled_tfrecord(example): UNLABELED_TFREC_FORMAT = { 'image': tf.io.FixedLenFeature([], tf.string), 'image_name': tf.io.FixedLenFeature([], tf.string) } example = tf.io.parse_single_example(example, UNLABELED_TFREC_FORMAT) return example['image'], example['image_name'] <count_values>
feature_cols = train_X.columns selector = SelectKBest(f_classif, k=high3) X_new = selector.fit_transform(train_X[feature_cols], train_y) selected_features = pd.DataFrame(selector.inverse_transform(X_new), index=train_X.index, columns=feature_cols) selected_cols1 = selected_features.columns[selected_features.var() !=...
Titanic - Machine Learning from Disaster
10,022,452
def count_data_items(filenames): n = [ int(re.compile(r'-([0-9]*)\.' ).search(filename ).group(1)) for filename in filenames ] return np.sum(n )<create_dataframe>
cv_high, cv_low = scores(cv_results )
Titanic - Machine Learning from Disaster
10,022,452
def getTrainDataset(files, cfg): ds = tf.data.TFRecordDataset(files, num_parallel_reads=AUTO) ds = ds.cache() opt = tf.data.Options() opt.experimental_deterministic = False ds = ds.with_options(opt) ds = ds.map(read_labeled_tfrecord, num_parallel_calls=AUTO) ds = ds.repeat() ds = ds.shuffle(2048) ds = ds.map(lambda...
cv_score = cross_val_score(best_k_model1, cv_X, cv_y, cv=cv_high) print('Mean cross-validation score: %.2f' %(cv_score.mean() *100))
Titanic - Machine Learning from Disaster
10,022,452
def getTestDataset(files, cfg, augment=False, repeat=False): ds = tf.data.TFRecordDataset(files, num_parallel_reads=AUTO) ds = ds.cache() if repeat: ds = ds.repeat() ds = ds.map(read_unlabeled_tfrecord, num_parallel_calls=AUTO) ds = ds.map(lambda img, idnum: (prepare_image(img, cfg=cfg), idnum), num_parallel_calls=A...
test_full.drop(columns=['Survived'], axis=0, inplace=True )
Titanic - Machine Learning from Disaster
10,022,452
def getLearnRateCallback(cfg): lr_start = cfg['lr_start'] lr_max = cfg['lr_max'] * strategy.num_replicas_in_sync * cfg['batch_size'] lr_min = cfg['lr_min'] lr_rampup = cfg['lr_rampup'] lr_sustain = cfg['lr_sustain'] lr_decay = cfg['lr_decay'] def lrfn(epoch): if epoch < lr_rampup: lr =(lr_max - lr_start)/ lr_rampup * e...
ktest_X = test_full[selected_cols] selected_cols
Titanic - Machine Learning from Disaster
10,022,452
with strategy.scope() : model_input = tf.keras.Input(shape=(cfg['img_size'], cfg['img_size'], 3), name='img_input') dummy = tf.keras.layers.Lambda(lambda x: x )(model_input) outputs = [] x = efn.EfficientNetB3(include_top=False, weights='imagenet', input_shape=(cfg['img_size'], cfg['img_size'], 3), pooling='avg' )(du...
pred = best_k_model.predict(ktest_X ).astype('int') pred
Titanic - Machine Learning from Disaster
10,022,452
ds_train = getTrainDataset(filenames_train, cfg ).map(lambda img, label:(img,(label, label, label))) stepsTrain = count_data_items(filenames_train)/(cfg['batch_size'] * strategy.num_replicas_in_sync )<train_model>
output = pd.DataFrame({'PassengerId': test_full.index, 'Survived': pred}) output.to_csv('survived.csv', index=False )
Titanic - Machine Learning from Disaster
10,022,452
history = model.fit(ds_train, validation_data = None, verbose=1, steps_per_epoch = stepsTrain, validation_steps = 0, epochs=14, callbacks=callbacks )<create_dataframe>
ktest_X1 = test_full[selected_cols1] pred1 = best_k_model1.predict(ktest_X1 ).astype('int') output = pd.DataFrame({'PassengerId': test_full.index, 'Survived': pred1}) output.to_csv('survived1.csv', index=False )
Titanic - Machine Learning from Disaster
10,022,452
<predict_on_test><EOS>
ktest_X2 = test_full[selected_cols2] pred2 = xg_best_model.predict(ktest_X2 ).astype('int') output = pd.DataFrame({'PassengerId': test_full.index, 'Survived': pred2}) output.to_csv('survived2.csv', index=False )
Titanic - Machine Learning from Disaster
5,201,782
<SOS> metric: categorizationaccuracy Kaggle data source: titanic-machine-learning-from-disaster<save_to_csv>
%matplotlib inline data_train = pd.read_csv('.. /input/train.csv') data_test = pd.read_csv('.. /input/test.csv' )
Titanic - Machine Learning from Disaster
5,201,782
y_test_sorted = np.zeros(( 3, probs.shape[1])) test = test.reset_index() test = test.set_index('image_name') i = 0 ds_test = getTestDataset(filenames_test, cfg) for img, imgid in tqdm(iter(ds_test.unbatch())) : imgid = imgid.numpy().decode('utf-8') y_test_sorted[:, test.loc[imgid]['index']] = probs[:, i, 0] i += 1 f...
data_train = data_train.drop(columns=['Name', 'Ticket', 'Fare', 'Cabin']) data_test = data_test.drop(columns=['Name', 'Ticket', 'Fare', 'Cabin'])
Titanic - Machine Learning from Disaster
5,201,782
!gzip submission_model_0.csv !gzip submission_model_1.csv !gzip submission_model_2.csv !gzip blended_effnets.csv<load_from_csv>
display(data_train.Age.value_counts(dropna=False ).sort_index()) display(data_test.Age.value_counts(dropna=False ).sort_index() )
Titanic - Machine Learning from Disaster
5,201,782
def MinMaxBestBaseStacking(input_folder, best_base, output_path): sub_base = pd.read_csv(best_base) all_files = os.listdir(input_folder) outs = [pd.read_csv(os.path.join(input_folder, f), index_col=0)for f in all_files] concat_sub = pd.concat(outs, axis=1) cols = list(map(lambda x: "target" + str(x), range(len(conca...
data_train.Age = data_train.Age.fillna(data_train.Age.mean()) data_test.Age = data_test.Age.fillna(data_test.Age.mean())
Titanic - Machine Learning from Disaster
5,201,782
MinMaxBestBaseStacking('.. /input/melanoma-ensemble-files/', '.. /input/melanoma-ensemble-files/blend_sub.csv', 'submission.csv' )<load_from_csv>
display(data_train.Embarked.value_counts(dropna=False)) display(data_test.Embarked.value_counts(dropna=False))
Titanic - Machine Learning from Disaster
5,201,782
!ls /kaggle/input/20201124-ensemble-1-testcsv<import_modules>
data_train.Embarked = data_train.Embarked.fillna('S')
Titanic - Machine Learning from Disaster
5,201,782
warnings.filterwarnings("ignore" )<import_modules>
b = data_train.pop('Survived') data_train = pd.concat([data_train, b], axis=1) display(data_train.head())
Titanic - Machine Learning from Disaster
5,201,782
l5kit.__version__<set_options>
display(data_train.Age.value_counts(dropna=False ).sort_index()) display(data_test.Age.value_counts(dropna=False ).sort_index() )
Titanic - Machine Learning from Disaster
5,201,782
def set_seed(seed): random.seed(seed) np.random.seed(seed) os.environ["PYTHONHASHSEED"] = str(seed) torch.manual_seed(seed) torch.cuda.manual_seed(seed) set_seed(42 )<init_hyperparams>
display(data_train.Sex.value_counts(dropna=False ).sort_index()) display(data_test.Sex.value_counts(dropna=False ).sort_index() )
Titanic - Machine Learning from Disaster
5,201,782
cfg = { 'format_version': 4, 'data_path': "/kaggle/input/lyft-motion-prediction-autonomous-vehicles", 'model_params': { 'model_architecture': 'resnet50', 'history_num_frames': 10, 'history_step_size': 1, 'history_delta_time': 0.1, 'future_num_frames': 50, 'future_step_size': 1, 'future_delta_time': 0.1, 'model_name': "...
display(( data_train.Pclass.value_counts(dropna=False ).sort_index())) display(( data_test.Pclass.value_counts(dropna=False ).sort_index()))
Titanic - Machine Learning from Disaster
5,201,782
DIR_INPUT = cfg["data_path"] os.environ["L5KIT_DATA_FOLDER"] = DIR_INPUT dm = LocalDataManager(None )<create_dataframe>
display(data_train.Survived.value_counts(dropna=False ).sort_index())
Titanic - Machine Learning from Disaster