kernel_id int64 24.2k 23.3M | prompt stringlengths 8 1.85M | completetion stringlengths 1 182k | comp_name stringlengths 5 57 |
|---|---|---|---|
14,068,986 | learning_rate = 0.001
def step_decay(epoch):
initial_lrate = learning_rate
drop = 0.1
epochs_drop = 20.0
lrate = initial_lrate * np.power(drop,
np.floor(( epoch)/epochs_drop))
tf.print("Learning rate: ", lrate)
return lrate
lrate = tf.keras.callbacks.LearningRateScheduler(step_decay)
early_stop = tf.keras.callbacks.E... | corpus_disaster, corpus_non_disaster = create_corpus(1), create_corpus(0)
counter_disaster, counter_non_disaster = Counter(corpus_disaster), Counter(corpus_non_disaster)
x_disaster, y_disaster, x_non_disaster, y_non_disaster = [], [], [], []
counter = 0
for word, count in counter_disaster.most_common() [0:100]:
if(wo... | Natural Language Processing with Disaster Tweets |
14,068,986 | mpnn = MPNN(mp_int_dim = 512, up_int_dim = 1024, out_int_dim = 512, state_dim = 256, T = 3)
mpnn.compile(opt, log_mae, metrics = [mae, log_mae])
<define_variables> | def bigrams(target):
corpus = train[train["target"] == target]["text"]
count_vec = CountVectorizer(ngram_range=(2, 2)).fit(corpus)
bag_of_words = count_vec.transform(corpus)
sum_words = bag_of_words.sum(axis=0)
words_freq = [(word, sum_words[0, idx])for word, idx in count_vec.vocabulary_.items() ]
words_freq =sorted... | Natural Language Processing with Disaster Tweets |
14,068,986 | batch_size = 64
epochs = 30
<train_model> | def remove_pattern(input_txt, pattern):
r = re.findall(pattern, input_txt)
for i in r:
input_txt = re.sub(i, '', input_txt)
return input_txt
train['tweet'] = np.vectorize(remove_pattern )(train['text'], "
test['tweet'] = np.vectorize(remove_pattern )(test['text'], "
train.head()
train['tweet'] = train['tweet'].str.re... | Natural Language Processing with Disaster Tweets |
14,068,986 | def train() :
nodes_train = np.load(datadir + "internalgraphdata/nodes_train.npz")['arr_0']
in_edges_train = np.load(datadir + "internalgraphdata/in_edges_train.npz")['arr_0']
out_edges_train = np.load(datadir + "internalgraphdata/out_edges_train.npz")['arr_0']
out_labels = out_edges_train.reshape(-1,out_edges_train.sh... | warnings.filterwarnings("ignore")
tqdm.pandas()
stopword=set(STOPWORDS)
lem = WordNetLemmatizer()
tokenizer=TweetTokenizer()
np.random.seed(0)
random_state = 29 | Natural Language Processing with Disaster Tweets |
14,068,986 | %%time
preds, train_size, history = train()<load_pretrained> | !pip install GPUtil
def free_gpu_cache() :
print("Initial GPU Usage")
gpu_usage()
torch.cuda.empty_cache()
cuda.select_device(0)
cuda.close()
cuda.select_device(0)
for obj in gc.get_objects() :
if torch.is_tensor(obj):
del obj
gc.collect()
print("GPU Usage after emptying the cache")
gpu_usage() | Natural Language Processing with Disaster Tweets |
14,068,986 | mpnn.save_weights("model.h5" )<load_pretrained> | train = pd.read_csv(".. /input/nlp-getting-started/train.csv")
test = pd.read_csv(".. /input/nlp-getting-started/test.csv")
sub= pd.read_csv(".. /input/nlp-getting-started/sample_submission.csv" ) | Natural Language Processing with Disaster Tweets |
14,068,986 | with open('/trainHistoryDict.pkl', 'wb')as file_pi:
pickle.dump(history.history, file_pi )<load_from_csv> | abbreviations = {
"$" : " dollar ",
"€" : " euro ",
"4ao" : "for adults only",
"a.m" : "before midday",
"a3" : "anytime anywhere anyplace",
"aamof" : "as a matter of fact",
"acct" : "account",
"adih" : "another day in hell",
"afaic" : "as far as i am concerned",
"afaict" : "as far as i can tell",
"afaik" : "as far as i... | Natural Language Processing with Disaster Tweets |
14,068,986 | train = pd.read_csv(datadir + "champs-scalar-coupling/train.csv")
test = pd.read_csv(datadir + "champs-scalar-coupling/test.csv")
train_mol_names = train['molecule_name'].unique()
val = train[train.molecule_name.isin(train_mol_names[train_size:])]
val_group = val.groupby('molecule_name' )<compute_test_metric> | def remove_URL(text):
url = re.compile(r'https?://\S+|www\.\S+')
return url.sub(r'URL',text)
def remove_HTML(text):
html=re.compile(r'<.*?>')
return html.sub(r'',text)
def remove_not_ASCII(text):
text = ''.join([word for word in text if word in string.printable])
return text
def word_abbrev(word):
return abbreviat... | Natural Language Processing with Disaster Tweets |
14,068,986 | def make_outs(test_group, preds):
i = 0
x = np.array([])
for test_gp, preds in zip(test_group, preds):
if(not i%1000):
print(i)
gp = test_gp[1]
x = np.append(x,(preds[gp['atom_index_0'].values, gp['atom_index_1'].values] + preds[gp['atom_index_1'].values, gp['atom_index_0'].values])/2.0)
i = i+1
return x
def group_m... | def clean_tweet(text):
text = remove_URL(text)
text = remove_HTML(text)
text = remove_not_ASCII(text)
text = text.lower()
text = replace_abbrev(text)
text = remove_mention(text)
text = remove_number(text)
text = remove_emoji(text)
text = transcription_sad(text)
text = transcription_smile(text)
text = transcrip... | Natural Language Processing with Disaster Tweets |
14,068,986 | max_size = 29
preds = preds.reshape(( -1,max_size, max_size))
out_unscaled = make_outs(val_group, preds )<feature_engineering> | train["clean_text"] = train["text"].apply(clean_tweet)
test["clean_text"] = test["text"].apply(clean_tweet)
train["clean_tokens"] = train["clean_text"].apply(lambda x: word_tokenize(x))
test["clean_tokens"] = test["clean_text"].apply(lambda x: word_tokenize(x)) | Natural Language Processing with Disaster Tweets |
14,068,986 | val['pred_scalar_coupling_constant'] = out_unscaled
coups_to_isolate = ['1JHC', '1JHN', '2JHC', '2JHH', '2JHN', '3JHC', '3JHH', '3JHN']
for i, coup in enumerate(coups_to_isolate):
scale_min = train['scalar_coupling_constant'].loc[train.type == coup].min()
scale_max = train['scalar_coupling_constant'].loc[train.type == ... | skip_gram_model = Word2Vec(train['clean_tokens'],size=150,window=3,min_count=2,sg=1)
skip_gram_model.train(train['clean_tokens'],total_examples=len(train['clean_tokens']),epochs=10)
cbow_model = Word2Vec(train['clean_tokens'],size=150,window=3,min_count=2)
cbow_model.train(train['clean_tokens'],total_examples=len(tr... | Natural Language Processing with Disaster Tweets |
14,068,986 | for coup in coups_to_isolate:
log_mae = group_mean_log_mae(val['scalar_coupling_constant'], val['pred_scalar_coupling_constant'], val['type'][val.type == coup])
print(coup,"\t", log_mae)
total = group_mean_log_mae(val['scalar_coupling_constant'], val['pred_scalar_coupling_constant'], val['type'])
print("")
print("T... | max_features=5000
count_vectorizer = CountVectorizer(max_features=max_features)
sparce_matrix_train=count_vectorizer.fit_transform(train['clean_text'])
sparce_matrix_test=count_vectorizer.fit_transform(train['clean_text'])
def count_vector(data):
count_vectorizer = CountVectorizer()
vect = count_vectorizer.fit_trans... | Natural Language Processing with Disaster Tweets |
14,068,986 | nodes_test = np.load(datadir + "internalgraphdata/nodes_test.npz")['arr_0']
in_edges_test = np.load(datadir + "internalgraphdata/in_edges_test.npz")['arr_0']
in_edges_test = in_edges_test.reshape(-1,in_edges_test.shape[1]*in_edges_test.shape[2],in_edges_test.shape[3] )<predict_on_test> | metrics = pd.DataFrame(columns=['model' ,'vectoriser', 'f1 score', 'train accuracy','test accuracy'] ) | Natural Language Processing with Disaster Tweets |
14,068,986 | preds = mpnn.predict({'adj_input' : in_edges_test, 'nod_input': nodes_test}, verbose=1 )<save_model> | models=[
XGBClassifier(max_depth=6, n_estimators=1000),
LogisticRegression(random_state=random_state),
SVC(random_state=random_state),
MultinomialNB() ,
DecisionTreeClassifier(random_state = random_state),
KNeighborsClassifier() ,
RandomForestClassifier(random_state=random_state),
] | Natural Language Processing with Disaster Tweets |
14,068,986 | np.save("preds_kernel.npy" , preds )<groupby> | for model in models:
y = train.target
x = X_train_count
x_train, x_test, y_train, y_test = train_test_split(x,y, test_size = 0.3)
fit_and_predict(model,x_train,x_test,y_train,y_test,'Count vector')
x = X_train_tfidf
x_train, x_test, y_train, y_test = train_test_split(x,y, test_size = 0.3)
fit_and_predict(model,x_tra... | Natural Language Processing with Disaster Tweets |
14,068,986 | test_group = test.groupby('molecule_name' )<normalization> | metrics = metrics.sort_values('f1 score',ascending=False ) | Natural Language Processing with Disaster Tweets |
14,068,986 | preds = preds.reshape(( -1,max_size, max_size))
out_unscaled = make_outs(test_group, preds )<feature_engineering> | free_gpu_cache() | Natural Language Processing with Disaster Tweets |
14,068,986 | test['scalar_coupling_constant'] = out_unscaled
coups_to_isolate = ['1JHC', '1JHN', '2JHC', '2JHH', '2JHN', '3JHC', '3JHH', '3JHN']
for i, coup in enumerate(coups_to_isolate):
scale_min = train['scalar_coupling_constant'].loc[train.type == coup].min()
scale_max = train['scalar_coupling_constant'].loc[train.type == coup... | from tensorflow.keras.preprocessing.text import Tokenizer
from tensorflow.keras.preprocessing.sequence import pad_sequences
from tensorflow import keras
from keras.models import Sequential
from keras.layers import Dense, Embedding, LSTM,GRU, Dropout, Activation, Input, Flatten, Bidirectional, Conv1D, MaxPooling1D
from ... | Natural Language Processing with Disaster Tweets |
14,068,986 | test[['id','scalar_coupling_constant']].to_csv('submission.csv', index=False )<install_modules> | def train_lstm(x_train,x_test,y_train,y_test,vectorizer_name,vocab_size,input_length):
epochs = 1
verbose = 1
batch_size = 32
embed_dim = 32
optimizer = optimizers.Adam(lr=0.002)
model = Sequential()
model.add(Embedding(vocab_size, embed_dim,input_length = input_length))
model.add(Dropout(0.2))
model.add(LSTM(32, drop... | Natural Language Processing with Disaster Tweets |
14,068,986 | !pip install tensorflow-gpu==2.0a0<import_modules> | y = train['target'].values
x_train, x_test, y_train, y_test = train_test_split(X_train_skip_gram,y, test_size = 0.3)
train_lstm(x_train,x_test,y_train,y_test, 'skip gram vector',5329,150)
| Natural Language Processing with Disaster Tweets |
14,068,986 | print(tf.__version__ )<set_options> | %reset -f | Natural Language Processing with Disaster Tweets |
14,068,986 | tf.test.is_gpu_available(
cuda_only=False,
min_cuda_compute_capability=None
)
<define_variables> | !pip install GPUtil
def free_gpu_cache() :
print("Initial GPU Usage")
gpu_usage()
torch.cuda.empty_cache()
cuda.select_device(0)
cuda.close()
cuda.select_device(0)
for obj in gc.get_objects() :
if torch.is_tensor(obj):
del obj
gc.collect()
print("GPU Usage after emptying the cache")
gpu_usage()
free_gpu_cache() | Natural Language Processing with Disaster Tweets |
14,068,986 | tf.random.set_seed(42)
datadir = ".. /input/"<choose_model_class> | import re
import torch
from transformers import ElectraTokenizer, ElectraForSequenceClassification,AdamW
import torch
from sklearn.metrics import classification_report
import random
import time
import datetime
import numpy as np
import pandas as pd
from transformers import get_linear_schedule_with_warmup
from torch.uti... | Natural Language Processing with Disaster Tweets |
14,068,986 | class Message_Passer_1(tf.keras.layers.Layer):
def __init__(self, intermediate_dim, state_dim):
super(Message_Passer_1, self ).__init__()
self.concat_layer = tf.keras.layers.Concatenate()
self.hidden_layer_1 = tf.keras.layers.Dense(units=intermediate_dim, activation=tf.nn.relu)
self.output_layer = tf.keras.layers.Dens... | if torch.cuda.is_available() :
device = torch.device("cuda")
print('We will use the GPU:', torch.cuda.get_device_name(0))
else:
print('No GPU available, using the CPU instead.')
device = torch.device("cpu" ) | Natural Language Processing with Disaster Tweets |
14,068,986 | class Message_Agg(tf.keras.layers.Layer):
def __init__(self):
super(Message_Agg, self ).__init__()
def call(self, messages):
return tf.math.reduce_sum(messages, 2 )<choose_model_class> | train = pd.read_csv(".. /input/nlp-getting-started/train.csv")
test = pd.read_csv(".. /input/nlp-getting-started/test.csv")
df_train= train
df_test= test | Natural Language Processing with Disaster Tweets |
14,068,986 | class Update_Func_1(tf.keras.layers.Layer):
def __init__(self, intermediate_dim, state_dim):
super(Update_Func_1, self ).__init__()
self.concat_layer = tf.keras.layers.Concatenate()
self.hidden_layer_1 = tf.keras.layers.Dense(units=intermediate_dim, activation=tf.nn.relu)
self.output_layer = tf.keras.layers.Dense(unit... | def preprocess(text):
text=text.lower()
text = re.sub(r'https?:\/\/.*[\r
]*', '', text)
text = re.sub(r'http?:\/\/.*[\r
]*', '', text)
text=text.replace(r'&?',r'and')
text=text.replace(r'<',r'<')
text=text.replace(r'>',r'>')
text = re.sub(r"(?:\@)\w+", '', text)
text=text.encode("ascii",errors="ignore" ... | Natural Language Processing with Disaster Tweets |
14,068,986 | class Adj_Updater_1(tf.keras.layers.Layer):
def __init__(self, intermediate_dim, state_dim):
super(Adj_Updater_1, self ).__init__()
self.concat_layer = tf.keras.layers.Concatenate()
self.hidden_layer_1 = tf.keras.layers.Dense(units=intermediate_dim, activation=tf.nn.relu)
self.output_layer = tf.keras.layers.Dense(unit... | df_train=df_train[["text","target"]] | Natural Language Processing with Disaster Tweets |
14,068,986 | class Edge_Regressor(tf.keras.layers.Layer):
def __init__(self, intermediate_dim):
super(Edge_Regressor, self ).__init__()
self.concat_layer = tf.keras.layers.Concatenate()
self.hidden_layer_1 = tf.keras.layers.Dense(units=intermediate_dim, activation=tf.nn.relu)
self.hidden_layer_2 = tf.keras.layers.Dense(units=inter... | texts = df_train.text.values
labels = df_train.target.values | Natural Language Processing with Disaster Tweets |
14,068,986 | class MP_Layer(tf.keras.layers.Layer):
def __init__(self, mp_int_dim, up_int_dim, out_int_dim, state_dim):
super(MP_Layer, self ).__init__(self)
self.state_dim = state_dim
self.message_passers = Message_Passer_1(intermediate_dim = mp_int_dim, state_dim = state_dim)
self.update_functions = Update_Func_1(intermediate_d... | torch.cuda.empty_cache()
tokenizer = ElectraTokenizer.from_pretrained('google/electra-base-discriminator')
model = ElectraForSequenceClassification.from_pretrained('google/electra-base-discriminator',num_labels=2)
model.cuda() | Natural Language Processing with Disaster Tweets |
14,068,986 | class MP_Layer_edge_only(tf.keras.layers.Layer):
def __init__(self, mp_int_dim, up_int_dim, out_int_dim, state_dim):
super(MP_Layer_edge_only, self ).__init__(self)
self.adj_updaters = Adj_Updater_1(intermediate_dim = up_int_dim, state_dim = state_dim)
self.message_aggs = Message_Agg()
self.state_dim = state_dim
def ... | indices=tokenizer.batch_encode_plus(texts,max_length=64,add_special_tokens=True, return_attention_mask=True,pad_to_max_length=True,truncation=True)
input_ids=indices["input_ids"]
attention_masks=indices["attention_mask"] | Natural Language Processing with Disaster Tweets |
14,068,986 | adj_input = tf.keras.Input(shape=(None,), name='adj_input')
nod_input = tf.keras.Input(shape=(None,), name='nod_input')
class MPNN(tf.keras.Model):
def __init__(self, mp_int_dim, up_int_dim, out_int_dim, state_dim, T):
super(MPNN, self ).__init__(self)
self.MP = [MP_Layer(mp_int_dim, up_int_dim, out_int_dim, state_d... | train_inputs, validation_inputs, train_labels, validation_labels = train_test_split(input_ids, labels,
random_state=42, test_size=0.2)
train_masks, validation_masks, _, _ = train_test_split(attention_masks, labels,
random_state=42, test_size=0.2 ) | Natural Language Processing with Disaster Tweets |
14,068,986 | def log_mae(orig , preds):
mask = tf.where(tf.equal(orig, 0), orig, tf.ones_like(orig))
nums = tf.boolean_mask(orig, mask)
preds = tf.boolean_mask(preds, mask)
reconstruction_error = tf.math.log(tf.reduce_mean(tf.abs(tf.subtract(nums, preds))))
return reconstruction_error
def mae(orig , preds):
mask = tf.where(tf.equ... | train_inputs = torch.tensor(train_inputs)
validation_inputs = torch.tensor(validation_inputs)
train_labels = torch.tensor(train_labels, dtype=torch.long)
validation_labels = torch.tensor(validation_labels, dtype=torch.long)
train_masks = torch.tensor(train_masks, dtype=torch.long)
validation_masks = torch.tensor(v... | Natural Language Processing with Disaster Tweets |
14,068,986 | mpnn = MPNN(mp_int_dim = 512, up_int_dim = 1024, out_int_dim = 512, state_dim = 256, T = 7)
mpnn.compile(opt, log_mae, metrics = [mae, log_mae])
<define_variables> | batch_size = 32
train_data = TensorDataset(train_inputs, train_masks, train_labels)
train_sampler = RandomSampler(train_data)
train_dataloader = DataLoader(train_data, sampler=train_sampler, batch_size=batch_size)
validation_data = TensorDataset(validation_inputs, validation_masks, validation_labels)
validation_sam... | Natural Language Processing with Disaster Tweets |
14,068,986 | batch_size = 64
epochs = 30
<train_model> | optimizer = AdamW(model.parameters() ,
lr = 6e-6,
eps = 1e-8
)
epochs = 5
total_steps = len(train_dataloader)* epochs
scheduler = get_linear_schedule_with_warmup(optimizer,
num_warmup_steps = 0,
num_training_steps = total_steps ) | Natural Language Processing with Disaster Tweets |
14,068,986 | def train() :
nodes_train = np.load(datadir + "internalgraphdata/nodes_train.npz")['arr_0']
in_edges_train = np.load(datadir + "internalgraphdata/in_edges_train.npz")['arr_0']
out_edges_train = np.load(datadir + "internalgraphdata/out_edges_train.npz")['arr_0']
out_labels = out_edges_train.reshape(-1,out_edges_train.sh... | def flat_accuracy(preds, labels):
pred_flat = np.argmax(preds, axis=1 ).flatten()
labels_flat = labels.flatten()
return np.sum(pred_flat == labels_flat)/ len(labels_flat ) | Natural Language Processing with Disaster Tweets |
14,068,986 | preds, train_size, history = train()<load_from_csv> | seed_val = 42
random.seed(seed_val)
np.random.seed(seed_val)
torch.manual_seed(seed_val)
torch.cuda.manual_seed_all(seed_val)
loss_values = []
for epoch_i in range(0, epochs):
print("")
print('======== Epoch {:} / {:} ========'.format(epoch_i + 1, epochs))
print('Training...')
t0 = time.time()
total_loss = 0
mode... | Natural Language Processing with Disaster Tweets |
14,068,986 | train = pd.read_csv(datadir + "champs-scalar-coupling/train.csv")
test = pd.read_csv(datadir + "champs-scalar-coupling/test.csv")
train_mol_names = train['molecule_name'].unique()
val = train[train.molecule_name.isin(train_mol_names[train_size:])]
val_group = val.groupby('molecule_name' )<compute_test_metric> | print("")
print("Running Validation...")
t0 = time.time()
model.eval()
preds=[]
true=[]
eval_loss, eval_accuracy = 0, 0
nb_eval_steps, nb_eval_examples = 0, 0
for batch in validation_dataloader:
batch = tuple(t.to(device)for t in batch)
b_input_ids, b_input_mask, b_labels = batch
with torch.no_grad() :
outputs = mod... | Natural Language Processing with Disaster Tweets |
14,068,986 | def make_outs(test_group, preds):
i = 0
x = np.array([])
for test_gp, preds in zip(test_group, preds):
if(not i%1000):
print(i)
gp = test_gp[1]
x = np.append(x,(preds[gp['atom_index_0'].values, gp['atom_index_1'].values] + preds[gp['atom_index_1'].values, gp['atom_index_0'].values])/2.0)
i = i+1
return x
def group_m... | flat_predictions = [item for sublist in preds for item in sublist]
flat_predictions = np.argmax(flat_predictions, axis=1 ).flatten()
flat_true_labels = [item for sublist in true for item in sublist] | Natural Language Processing with Disaster Tweets |
14,068,986 | max_size = 29
preds = preds.reshape(( -1,max_size, max_size))
out_unscaled = make_outs(val_group, preds )<feature_engineering> | comments1 = df_test.text.values
indices1=tokenizer.batch_encode_plus(comments1,max_length=128,add_special_tokens=True, return_attention_mask=True,pad_to_max_length=True,truncation=True)
input_ids1=indices1["input_ids"]
attention_masks1=indices1["attention_mask"]
prediction_inputs1= torch.tensor(input_ids1)
prediction... | Natural Language Processing with Disaster Tweets |
14,068,986 | val['pred_scalar_coupling_constant'] = out_unscaled
coups_to_isolate = ['1JHC', '1JHN', '2JHC', '2JHH', '2JHN', '3JHC', '3JHH', '3JHN']
for i, coup in enumerate(coups_to_isolate):
scale_min = train['scalar_coupling_constant'].loc[train.type == coup].min()
scale_max = train['scalar_coupling_constant'].loc[train.type == ... | print('Predicting labels for {:,} test sentences...'.format(len(prediction_inputs1)))
model.eval()
predictions = []
for batch in prediction_dataloader1:
batch = tuple(t.to(device)for t in batch)
b_input_ids1, b_input_mask1 = batch
with torch.no_grad() :
outputs1 = model(b_input_ids1, token_type_ids=None,
attention_ma... | Natural Language Processing with Disaster Tweets |
14,068,986 | for coup in coups_to_isolate:
log_mae = group_mean_log_mae(val['scalar_coupling_constant'], val['pred_scalar_coupling_constant'], val['type'][val.type == coup])
print(coup,"\t", log_mae)
total = group_mean_log_mae(val['scalar_coupling_constant'], val['pred_scalar_coupling_constant'], val['type'])
print("")
print("T... | sample_sub=pd.read_csv('.. /input/nlp-getting-started/sample_submission.csv')
submit=pd.DataFrame({'id':sample_sub['id'].values.tolist() ,'target':flat_predictions} ) | Natural Language Processing with Disaster Tweets |
14,068,986 | <predict_on_test><EOS> | df_leak = pd.read_csv('/kaggle/input/disasters-on-social-media/socialmedia-disaster-tweets-DFE.csv', encoding ='ISO-8859-1')[['choose_one', 'text']]
df_leak['target'] =(df_leak['choose_one'] == 'Relevant' ).astype(np.int8)
df_leak['id'] = df_leak.index.astype(np.int16)
df_leak.drop(columns=['choose_one', 'text'], inp... | Natural Language Processing with Disaster Tweets |
13,985,799 | <SOS> metric: meanfscore Kaggle data source: natural-language-processing-with-disaster-tweets<save_model> | sns.set_style("darkgrid")
| Natural Language Processing with Disaster Tweets |
13,985,799 | np.save("preds_kernel.npy" , preds )<groupby> | df_train = pd.read_csv("/kaggle/input/nlp-getting-started/train.csv")
print(df_train.shape)
df_train.head() | Natural Language Processing with Disaster Tweets |
13,985,799 | test_group = test.groupby('molecule_name' )<normalization> | df_test = pd.read_csv("/kaggle/input/nlp-getting-started/test.csv")
print(df_test.shape)
df_test.head() | Natural Language Processing with Disaster Tweets |
13,985,799 | preds = preds.reshape(( -1,max_size, max_size))
out_unscaled = make_outs(test_group, preds )<feature_engineering> | def remove_emoji(text):
emoji_pattern = re.compile("["
u"\U0001F600-\U0001F64F"
u"\U0001F300-\U0001F5FF"
u"\U0001F680-\U0001F6FF"
u"\U0001F1E0-\U0001F1FF"
u"\U00002702-\U000027B0"
u"\U000024C2-\U0001F251"
"]+", flags=re.UNICODE)
return emoji_pattern.sub(r'', text ) | Natural Language Processing with Disaster Tweets |
13,985,799 | test['scalar_coupling_constant'] = out_unscaled
coups_to_isolate = ['1JHC', '1JHN', '2JHC', '2JHH', '2JHN', '3JHC', '3JHH', '3JHN']
for i, coup in enumerate(coups_to_isolate):
scale_min = train['scalar_coupling_constant'].loc[train.type == coup].min()
scale_max = train['scalar_coupling_constant'].loc[train.type == coup... | set_stopwords = set(stopwords.words('english'))
df_train['text_processed'] = df_train['text'].apply(lambda x: re.compile(r'https?://\S+|www\.\S+' ).sub(r'',x))
df_test['text_processed'] = df_test['text'].apply(lambda x: re.compile(r'https?://\S+|www\.\S+' ).sub(r'',x))
df_train['text_processed'] = df_train['text_proces... | Natural Language Processing with Disaster Tweets |
13,985,799 | test[['id','scalar_coupling_constant']].to_csv('submission.csv', index=False )<define_variables> | abbreviations = {
"$" : " dollar ",
"€" : " euro ",
"4ao" : "for adults only",
"a.m" : "before midday",
"a3" : "anytime anywhere anyplace",
"aamof" : "as a matter of fact",
"acct" : "account",
"adih" : "another day in hell",
"afaic" : "as far as i am concerned",
"afaict" : "as far as i can tell",
"afaik" : "as far as i... | Natural Language Processing with Disaster Tweets |
13,985,799 | dtypes = {'atom_index_0':'uint8',
'atom_index_1':'uint8',
'scalar_coupling_constant':'float32',
'num_C':'uint8',
'num_H':'uint8',
'num_N':'uint8',
'num_O':'uint8',
'num_F':'uint8',
'total_atoms':'uint8',
'num_bonds':'uint8',
'num_mol_bonds':'uint8',
'min_d':'float32',
'mean_d':'float32',
'max_d':'float32',
'space_dr':'... | def convert_abbrev(word):
return abbreviations[word.lower() ] if word.lower() in abbreviations.keys() else word
df_train['text_processed'] = df_train['text_processed'].apply(lambda x: ' '.join([convert_abbrev(word)for word in word_tokenize(x)]))
df_test['text_processed'] = df_test['text_processed'].apply(lambda x: ' '.... | Natural Language Processing with Disaster Tweets |
13,985,799 | train = pd.read_csv(".. /input/predmolprop-featureengineering-final/train_extend.csv",dtype=dtypes)
test = pd.read_csv(".. /input/predmolprop-featureengineering-finaltest/test_extend.csv",dtype=dtypes )<categorify> | ids_with_target_error = [328,443,513,2619,3640,3900,4342,5781,6552,6554,6570,6701,6702,6729,6861,7226]
df_train.loc[df_train['id'].isin(ids_with_target_error),'target'] = 0 | Natural Language Processing with Disaster Tweets |
13,985,799 | cols = ['atom_0_type2','atom_2_type','atom_3_type','atom_end_type2']
for col in cols:
enc = LabelEncoder()
train[col]=enc.fit_transform(train[col] ).astype(np.uint8)
test[col]=enc.transform(test[col] ).astype(np.uint8)
del cols<define_variables> | learning_rate = 1e-5
valid = 0.2
epochs_num = 3
batch_size_num = 16 | Natural Language Processing with Disaster Tweets |
13,985,799 | prefix_train= ['id', 'type', 'scalar_coupling_constant']
prefix_test= ['id', 'type']
fc_distance = ['space_dr','min_d','mean_d', 'max_d']
fc_COM = ['Dmin_COM', 'Dmean_COM', 'Dmax_COM']
fc_size = ['num_mol_bonds', 'total_atoms','num_C', 'num_H', 'num_N', 'num_O', 'num_F']
fc_atom_0 = ['atom_0_pc','atom_0_type2','COM_dr_... | !wget --quiet https://raw.githubusercontent.com/tensorflow/models/master/official/nlp/bert/tokenization.py | Natural Language Processing with Disaster Tweets |
13,985,799 | def ProcessData(df,features,test_size=0.25):
if test_size == 0:
train_Y = df.pop('scalar_coupling_constant')
train_type = df.pop('type')
df.pop('id')
return df.loc[:,df.columns.map(lambda x: x in features)], train_Y, train_type
train_X, val_X, train_Y, val_Y = train_test_split(df.loc[:,df.columns.map(lambda x: x in ... | import tensorflow as tf
from tensorflow.keras.layers import Dense, Input
from tensorflow.keras.optimizers import Adam
from tensorflow.keras.models import Model
from tensorflow.keras.callbacks import ModelCheckpoint
import tensorflow_hub as hub
import tokenization | Natural Language Processing with Disaster Tweets |
13,985,799 | def CalcLMAE(y_true, y_pred, groups, floor=1e-9):
maes =(y_true-y_pred ).abs().groupby(groups ).mean()
return np.log(maes.map(lambda x: max(x, floor)) ).mean()<categorify> | def bert_encode(texts, tokenizer, max_len=512):
all_tokens = []
all_masks = []
all_segments = []
for text in texts:
text = tokenizer.tokenize(text)
text = text[:max_len-2]
input_sequence = ["[CLS]"] + text + ["[SEP]"]
pad_len = max_len - len(input_sequence)
tokens = tokenizer.convert_tokens_to_ids(input_sequence)
to... | Natural Language Processing with Disaster Tweets |
13,985,799 | def SingleRun(df,features, test_size=0.25, model_fn=XGBRegressor, includeType=False, early_stopping_rounds=None, do_SHAP=False, **kwargs):
data = ProcessData(df,features,test_size)
if(test_size==0):
train_X,train_Y,train_type = data
else:
train_X,train_Y,train_type,val_X,val_Y,val_type = data
if includeType:
train_X=t... | def build_model(bert_layer, max_len=512):
input_word_ids = Input(shape=(max_len,), dtype=tf.int32, name="input_word_ids")
input_mask = Input(shape=(max_len,), dtype=tf.int32, name="input_mask")
segment_ids = Input(shape=(max_len,), dtype=tf.int32, name="segment_ids")
_, sequence_output = bert_layer([input_word_ids, ... | Natural Language Processing with Disaster Tweets |
13,985,799 | coupling_type = '3JHN'
train_sample = train[train.type==coupling_type]
features=fc1.copy()
if(coupling_type[0]=='2'):
features.update(fc_2)
elif(coupling_type[0]=='3'):
features.update(fc_2[:-1]+fc_3)
model_3JHN,_,_=SingleRun(train_sample,features,test_size=0.2,model_fn=XGBRegressor,includeType=False,early_stopping_r... | module_url = "https://tfhub.dev/tensorflow/bert_en_uncased_L-24_H-1024_A-16/1"
bert_layer = hub.KerasLayer(module_url, trainable=True ) | Natural Language Processing with Disaster Tweets |
13,985,799 | def RunByType(df,test_size=0.25,model_fn=XGBRegressor,includeType=False,early_stopping_rounds=None,**kwargs):
model_dict={}
train_LMAE_dict={}
val_LMAE_dict={}
for coupling_type in coupling_types:
print('Now training type:',str(coupling_type))
df_type = df[df['type']==coupling_type]
features=fc1.copy()
if(coupling_type... | vocab_file = bert_layer.resolved_object.vocab_file.asset_path.numpy()
do_lower_case = bert_layer.resolved_object.do_lower_case.numpy()
tokenizer = tokenization.FullTokenizer(vocab_file, do_lower_case ) | Natural Language Processing with Disaster Tweets |
13,985,799 | def PredictByType(df_X,model_dict,df_Y=None):
predictions = pd.DataFrame()
for coupling_type in coupling_types:
print('predicting type:',str(coupling_type))
model = model_dict[coupling_type]
df_type = df_X[df_X['type']==coupling_type]
features=fc1.copy()
if(coupling_type[0]=='2'):
features.update(fc_2)
elif(coupling_t... | train_input = bert_encode(df_train['text_processed'].values, tokenizer, max_len=160)
test_input = bert_encode(df_test['text_processed'].values, tokenizer, max_len=160)
train_labels = df_train['target'].values | Natural Language Processing with Disaster Tweets |
13,985,799 | model_dict=RunByType(train,test_size=0.2,includeType=False,early_stopping_rounds=5,
max_depth=11, learning_rate=0.1, n_estimators=10000,
verbosity=1,
objective='reg:squarederror', booster='gbtree',tree_method= 'gpu_hist',
n_jobs=4,
gamma=0, min_child_weight=1, max_delta_step=0,
subsample=1,colsample_bytree=1, colsample... | model_BERT = build_model(bert_layer, max_len=160)
model_BERT.summary() | Natural Language Processing with Disaster Tweets |
13,985,799 | print('FINISHED!' )<load_from_csv> | checkpoint = ModelCheckpoint('model_BERT.h5', monitor='val_loss', save_best_only=True)
train_history = model_BERT.fit(
train_input, train_labels,
validation_split = valid,
epochs = epochs_num,
callbacks=[checkpoint],
batch_size = batch_size_num
) | Natural Language Processing with Disaster Tweets |
13,985,799 | one = pd.read_csv('.. /input/champs-blending-tutorial/1.csv')
two = pd.read_csv('.. /input/champs-blending-tutorial/2.csv')
three = pd.read_csv('.. /input/champs-blending-tutorial/3.csv')
four = pd.read_csv('.. /input/otherkernelsadded/submission-2.csv')
five = pd.read_csv('.. /input/otherkernelsadded/submission-gi... | test_pred = model_BERT.predict(test_input)
test_pred_int = test_pred.round().astype('int')
train_pred = model_BERT.predict(train_input)
train_pred_int = train_pred.round().astype('int' ) | Natural Language Processing with Disaster Tweets |
13,985,799 | warnings.filterwarnings("ignore")
warnings.filterwarnings(action="ignore",category=DeprecationWarning)
warnings.filterwarnings(action="ignore",category=FutureWarning )<compute_train_metric> | print("F1 Score = " + str(f1_score(df_train['target'], train_pred_int)) ) | Natural Language Processing with Disaster Tweets |
13,985,799 | <load_from_csv><EOS> | df_submission = pd.read_csv("/kaggle/input/nlp-getting-started/sample_submission.csv")
df_submission['target'] = test_pred_int
df_submission.to_csv("submission.csv", index=False, header=True ) | Natural Language Processing with Disaster Tweets |
13,104,084 | <SOS> metric: meanfscore Kaggle data source: natural-language-processing-with-disaster-tweets<load_from_csv> | import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import nltk
from nltk.corpus import stopwords
import tensorflow as tf
import tensorflow_addons as tfa
from transformers import TFAutoModel, AutoTokenizer | Natural Language Processing with Disaster Tweets |
13,104,084 | scores_nn = dict()
for mol_type_index, mol_type in enumerate(mol_types):
print(mol_type, f'- run number {run_number}')
try:
scores_nn = np.load(f'run_{run_number}_scores_nn.npy', allow_pickle=True ).item()
except:
scores_nn = dict()
train = pd.read_csv(train_and_test_with_feats_folder + '/train_' + mol_type + '.csv' )... | tweet_tokenizer = TweetTokenizer()
def normalizeToken(token):
lowercased_token = token.lower()
if token.startswith("@"):
return "@USER"
elif lowercased_token.startswith("http")or lowercased_token.startswith("www"):
return "HTTPURL"
elif len(token)== 1:
return demojize(token)
else:
if token == "’":
return "'"
elif toke... | Natural Language Processing with Disaster Tweets |
13,104,084 | sub = pd.read_csv(f'{preds_and_oofs_folder}/final_model_submission.csv')
sub.to_csv('final_sub.csv', index=False )<import_modules> | def load_train_set() :
df = pd.read_csv("/kaggle/input/nlp-getting-started/train.csv")[["text", "target"]]
df["text"] = df["text"].apply(normalizeTweet)
return df
def load_test_set() :
df = pd.read_csv("/kaggle/input/nlp-getting-started/test.csv")[["id", "text"]]
df["text"] = df["text"].apply(normalizeTweet)
return d... | Natural Language Processing with Disaster Tweets |
13,104,084 | print(os.listdir('.. /input/'))
<load_from_csv> | print(train['target'].value_counts())
print()
print(train['target'].value_counts(normalize=True)) | Natural Language Processing with Disaster Tweets |
13,104,084 | %%time
def group_mean_log_mae(y_true, y_pred, types, floor=1e-9):
maes =(y_true-y_pred ).abs().groupby(types ).mean()
return np.log(maes.map(lambda x: max(x, floor)) ).mean()
train = pd.read_csv('.. /input/pmp-oof/final_train_oof_pmp.csv')
test = pd.read_csv('.. /input/pmp-oof/final_test_oof_pmp.csv')
drop_features... | disaster_tweets = train[train['target']==1]['text']
non_disaster_tweets = train[train['target']==0]['text']
freq_dist_disaster_tweets= nltk.FreqDist([word for tweet in disaster_tweets for word in tweet.lower().split() if word not in stopwords.words("english")and len(word)> 2])
freq_dist_non_disaster_tweets= nltk.FreqD... | Natural Language Processing with Disaster Tweets |
13,104,084 | def get_median_from_files(files):
print(len(files))
outs = [pd.read_csv(f, index_col=0)for f in files]
concat_sub = pd.concat(outs, axis=1, sort=True)
champ_median = concat_sub.median(axis=1 ).values
return champ_median
test = pd.read_csv(f".. /input/champs-scalar-coupling/test.csv")
TARGET = 'scalar_coupling_constan... | MAX_LENGTH = 50
short_tweets = sum(np.array(tweets_length)<= MAX_LENGTH)
long_tweets = sum(np.array(tweets_length)> MAX_LENGTH)
print("{} reviews with LEN > {}({:.2f} % of total data)".format(
long_tweets,
MAX_LENGTH,
100 * long_tweets / len(train)
)) | Natural Language Processing with Disaster Tweets |
13,104,084 | %matplotlib inline
test['nnet_ens'] = test['nnet_cont'] * 0.6 + test['nnet'] * 0.4
test['lgb_ens'] = test['lgb_a'] * 0.8 + test['lgb_m'] * 0.2
test['final_preds'] =(
test['n1']*0.5 +
test['n2']*0.1 +
test['lgb_ens']*0.15 +
test['nnet_ens']*0.1 +
test['lb']*0.1 +
test['final_mpnn'] * 0.03 + test['mpnn'] *0.02
)
test.... | def encode_tweets(tokenizer, tweets, max_len):
nb_tweets = len(tweets)
tokens = np.ones(( nb_tweets,max_len),dtype='int32')
masks = np.zeros(( nb_tweets,max_len),dtype='int32')
segs = np.zeros(( nb_tweets,max_len),dtype='int32')
for k in range(nb_tweets):
tweet = tweets[k]
enc = tokenizer.encode(tweet)
if len(enc)... | Natural Language Processing with Disaster Tweets |
13,104,084 | submission = pd.DataFrame()
submission['id'] = test.id
submission['scalar_coupling_constant'] = test['final_preds'] * 0.1 + test['stack15'] * 0.9
submission.to_csv('ensemble_sub.csv', index=False )<set_options> | train_tokens, train_masks, train_segs = encode_tweets(tokenizer,train["text"].to_list() , MAX_LENGTH)
train_labels = train["target"] | Natural Language Processing with Disaster Tweets |
13,104,084 | print(pd.__version__)
SEED = 26
LR = 1e-4<set_options> | es = tf.keras.callbacks.EarlyStopping(monitor='val_loss', mode='min', patience=3, restore_best_weights=True, verbose=1)
train_labels = train['target']
train_history = model.fit(
[train_tokens,train_masks,train_segs], train_labels,
validation_split=0.2,
epochs=5,
batch_size=16,
verbose = 1,
callbacks = [es]
) | Natural Language Processing with Disaster Tweets |
13,104,084 | <define_variables><EOS> | test_tokens, test_masks, test_segs = encode_tweets(tokenizer,test["text"].to_list() , MAX_LENGTH)
test["target"] = model.predict([test_tokens, test_masks, test_segs] ).round().astype(int)
submission = test[["id", "target"]]
submission.to_csv("submission.csv",index=False ) | Natural Language Processing with Disaster Tweets |
12,981,393 | <SOS> metric: meanfscore Kaggle data source: natural-language-processing-with-disaster-tweets<define_variables> | for dirname, _, filenames in os.walk('/kaggle/working'):
for filename in filenames:
print(os.path.join(dirname, filename))
print('นำเข้าไลบรารี่ข้อมูลเรียบร้อย' ) | Natural Language Processing with Disaster Tweets |
12,981,393 | types_dict = {
"1JHC": 0,
"2JHH": 3,
"1JHN": 1,
"2JHN": 4,
"2JHC": 2,
"3JHH": 6,
"3JHC": 5,
"3JHN": 7,
}
atom_features = [
"atom_2",
"atom_3",
"atom_4",
"atom_5",
"atom_6",
"atom_7",
"atom_8",
"atom_9",
]
fc_feats = [
"fc_preds_type",
"fc_preds_akira",
"fc_preds_akira2",
"fc_preds_akira3",
"fc_preds_akira4",
]
types_to... | train = pd.read_csv(".. /input/nlp-getting-started/train.csv")
test = pd.read_csv(".. /input/nlp-getting-started/test.csv")
sample_submission = pd.read_csv('/kaggle/input/nlp-getting-started/sample_submission.csv')
print('นำเข้าชุดข้อมูลเรียบร้อย' ) | Natural Language Processing with Disaster Tweets |
12,981,393 | train_dtypes = {
'molecule_name': 'category',
'atom_index_0': 'int8',
'atom_index_1': 'int8',
'type': 'category',
'scalar_coupling_constant': 'float32'
}
train_csv = pd.read_csv(f'{DATA_PATH}/train.csv', index_col='id', dtype=train_dtypes)
cols = ['molecule_name', 'atom_index_0', 'atom_index_1', 'type']
train_csv = tr... | print(train.isnull().sum() ) | Natural Language Processing with Disaster Tweets |
12,981,393 | test_csv = pd.read_csv(f'{DATA_PATH}/test.csv', index_col='id', dtype=train_dtypes)
test_csv['molecule_index'] = test_csv['molecule_name'].str.replace('dsgdb9nsd_', '' ).astype('int32')
cols = [col for col in cols if 'scalar_coupling_constant' not in col]
test_csv = test_csv[cols]
gc.collect()
print(train_csv.shape)
... | print(test.isnull().sum() ) | Natural Language Processing with Disaster Tweets |
12,981,393 | train_csv, test_csv = add_contributions(train_csv, test_csv )<drop_column> | print(sample_submission.isnull().sum() ) | Natural Language Processing with Disaster Tweets |
12,981,393 | cols = ["molecule_name", "atom_index_0", "atom_index_1"]
train_csv = train_csv.drop(cols, axis=1)
test_csv = test_csv.drop(cols, axis=1)
train_csv.head()<load_from_csv> | !wget --quiet https://raw.githubusercontent.com/tensorflow/models/master/official/nlp/bert/tokenization.py
print('ดาวน์โหลด Algorithm สำเร็จ!!!' ) | Natural Language Processing with Disaster Tweets |
12,981,393 | tr = read_pickle(FILETRAIN)
tr = tr.fillna(0)
train_ix = train_csv.index
tr.index = train_ix
train_csv = pd.concat([train_csv, tr], axis=1)
train_csv.index = train_ix
train_csv = train_csv[[col for col in train_csv.columns if col in list(all_feats)+ TARGETS + ["type"]]]
del tr<load_from_csv> | print('นำเข้า "tokenization" เรียบร้อย' ) | Natural Language Processing with Disaster Tweets |
12,981,393 | te = read_pickle(FILETEST)
te = te.fillna(0)
test_ix = test_csv.index
te.index = test_ix
test_csv = pd.concat([test_csv, te], axis=1)
test_csv.index = test_ix
test_csv = test_csv[[col for col in test_csv.columns if col in list(all_feats)+ ["type"]]]
del te, test_ix, train_ix
gc.collect()
print(train_csv.shape)
prin... | def bert_encode(texts, tokenizer, max_len=512):
all_tokens = []
all_masks = []
all_segments = []
for text in texts:
text = tokenizer.tokenize(text)
text = text[:max_len-2]
input_sequence = ["[CLS]"] + text + ["[SEP]"]
pad_len = max_len - len(input_sequence)
tokens = tokenizer.convert_tokens_to_ids(input_sequence)... | Natural Language Processing with Disaster Tweets |
12,981,393 | tr = read_pickle(FILETRAIN1)
tr = tr.fillna(0)[[col for col in tr.columns if col not in train_csv.columns]]
train_ix = train_csv.index
tr.index = train_ix
train_csv = pd.concat([train_csv, tr], axis=1)
train_csv.index = train_ix
del tr
train_csv = train_csv[list(all_feats)+ TARGETS + ["type"]]
gc.collect()
te = read_... | "Let's learn deep learning!"
['Let', "'", 's', 'learn', 'deep', 'learning', '!']
['[CLS]', 'Let', "'", 's', 'learn', 'deep', 'learning', '!', '[SEP]']
['[CLS]', 'Let', "'", 's', 'learn', 'deep', 'learning', '!', '[SEP]', '[PAD]','[PAD]','[PAD]','[PAD]','[PAD]']
[101, 2421, 112, 188, 3858, 1996, 3776, 106, 102, 0, 0, 0,... | Natural Language Processing with Disaster Tweets |
12,981,393 | gc.collect()
print(len(train_csv.columns), len(np.unique(train_csv.columns)) )<set_options> | def build_model(bert_layer, max_len=512):
input_word_ids = Input(shape=(max_len,), dtype=tf.int32, name="input_word_ids")
input_mask = Input(shape=(max_len,), dtype=tf.int32, name="input_mask")
segment_ids = Input(shape=(max_len,), dtype=tf.int32, name="segment_ids")
_, sequence_output = bert_layer([input_word_ids, ... | Natural Language Processing with Disaster Tweets |
12,981,393 | config = tf.ConfigProto(device_count = {'GPU': 1 , 'CPU': 2})
config.gpu_options.allow_growth = True
config.gpu_options.per_process_gpu_memory_fraction = 0.6
sess = tf.Session(config=config)
K.set_session(sess )<categorify> | %%time
print('กำลังดาวน์โหลดโมเดลอาจใช้เวลาสักครู่...')
module_url = "https://tfhub.dev/tensorflow/bert_en_uncased_L-24_H-1024_A-16/1"
bert_layer = hub.KerasLayer(module_url, trainable=True)
print('ดาวน์โหลดโมเดลสำเร็จ!!!' ) | Natural Language Processing with Disaster Tweets |
12,981,393 | cv_score = []
cv_score_total = 0
retrain = True
start_time = datetime.now()
test_prediction = np.zeros(len(test_csv))
class FeatureTransformer:
def transform(self, dataset, ohe_features=[], continuous_features=[]):
ohe_df = OneHotEncoder().fit_transform(dataset.loc[:, ohe_features] ).toarray()
skews = dataset.loc[:, co... | vocab_file = bert_layer.resolved_object.vocab_file.asset_path.numpy()
do_lower_case = bert_layer.resolved_object.do_lower_case.numpy()
tokenizer = tokenization.FullTokenizer(vocab_file, do_lower_case ) | Natural Language Processing with Disaster Tweets |
12,981,393 | submit = pd.read_csv(f'{DATA_PATH}/sample_submission.csv')
def submits(predictions):
submit["scalar_coupling_constant"] = predictions
submit.to_csv("/kaggle/working/nnetCont_sub.csv", index=False)
submits(test_prediction )<train_model> | train_input = bert_encode(train.text.values, tokenizer, max_len=160)
test_input = bert_encode(test.text.values, tokenizer, max_len=160)
train_labels = train.target.values | Natural Language Processing with Disaster Tweets |
12,981,393 | print('Total training time: ', datetime.now() - start_time)
i=0
for mol_type in types_to_run:
print(mol_type,": cv score is ",cv_score[i])
i+=1
print("total cv score is",cv_score_total )<save_to_csv> | start_time = time.time()
train_history = model.fit(train_input,
train_labels,
validation_split = 0.2,
epochs = 3,
batch_size = 16)
print(' สำเร็จ!!! ')
print("---ใช้เวลาทั้งหมด %s วินาที ---" %(time.time() - start_time)) | Natural Language Processing with Disaster Tweets |
12,981,393 | submit = pd.read_csv(f'{DATA_PATH}/sample_submission.csv')
def submits(predictions):
submit["scalar_coupling_constant"] = predictions
submit.to_csv(f"/kaggle/working/nnetCont_sub_{round(cv_score_total, 4)}.csv", index=False)
submits(test_prediction )<set_options> | test_pred = model.predict(test_input)
print(' สำเร็จ!!! ' ) | Natural Language Processing with Disaster Tweets |
12,981,393 | %%capture
warnings.filterwarnings('ignore')
DATA_DIR = '.. /input/champs-scalar-coupling'
ATOMIC_NUMBERS = {
'H': 1,
'C': 6,
'N': 7,
'O': 8,
'F': 9
}<load_from_csv> | sample_submission['target'] = test_pred.round().astype(int)
sample_submission.to_csv('sample_submission.csv', index=False)
print('สร้างไฟล์ submission.csv เรียบร้อย' ) | Natural Language Processing with Disaster Tweets |
12,981,393 | <merge><EOS> | sample_submission.isnull().sum() | Natural Language Processing with Disaster Tweets |
13,037,893 | <SOS> metric: meanfscore Kaggle data source: natural-language-processing-with-disaster-tweets<load_from_csv> | import numpy as np
import pandas as pd
from sklearn.feature_extraction.text import TfidfVectorizer ,TfidfTransformer | Natural Language Processing with Disaster Tweets |
13,037,893 | def load_j_coupling_csv(file_path: str, train=True, verbose=False):
train_dtypes = {
'molecule_name': 'category',
'atom_index_0': 'int8',
'atom_index_1': 'int8',
'type': 'category',
'scalar_coupling_constant': 'float32'
}
df = pd.read_csv(file_path, dtype=train_dtypes)
df['molecule_index'] = df.molecule_name.str.repla... | nlp = spacy.load("en_core_web_lg" ) | Natural Language Processing with Disaster Tweets |
13,037,893 | def get_knn_features_center(j_coupling: pd.Series,
structures=structures_df,
mol2dist=mol2distance_matrix,
k=10)-> np.array:
center = j_coupling[['x_c', 'y_c', 'z_c']].values.reshape(1, 3)
mol_df = structures.loc[j_coupling.molecule_index]
coordinates = mol_df[['x','y', 'z']].values
center_distances = distance_matrix(... | train_data = pd.read_csv('.. /input/nlp-getting-started/train.csv')
test_data =pd.read_csv('.. /input/nlp-getting-started/test.csv')
train_data.head(5 ) | Natural Language Processing with Disaster Tweets |
13,037,893 | def make_data(df: pd.DataFrame, id2features: dict, random_state=128, split=True):
tmp_df = df.copy()
tmp_df['features'] = tmp_df.id.map(id2features)
X = np.stack(tmp_df.features)
y = tmp_df.scalar_coupling_constant.values
if split:
X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=0.2, random_state=ra... | train_data_shape = train_data.shape[0] | Natural Language Processing with Disaster Tweets |
13,037,893 | print(f'competition-metric: {np.mean(list(scores.values())) :.2f}')
print('scores per type:')
pprint(scores, width=1 )<prepare_x_and_y> | def clean_text(text):
url = re.compile(r'https?://\S+|www\.\S+')
text = url.sub(r'', text)
html = re.compile(r'<.*?>')
text = html.sub(r'', text)
emoji_pattern = re.compile("["
u"\U0001F600-\U0001F64F"
u"\U0001F300-\U0001F5FF"
u"\U0001F680-\U0001F6FF"
u"\U0001F1E0-\U0001F1FF"
u"\U00002702-\U000027B0"
u"\U000024C2-\... | Natural Language Processing with Disaster Tweets |
13,037,893 | def make_test_data(df: pd.DataFrame, id2features: dict, random_state=128):
tmp_df = df.copy()
tmp_df['features'] = tmp_df.id.map(id2features)
X = np.stack(tmp_df.features)
return X
test_df = load_j_coupling_csv(join(DATA_DIR, 'test.csv'), train=False, verbose=True)
id2center_knn_test = {row.id : get_knn_features_cen... | def massage_text(text):
tweet = re.sub("[^a-zA-Z]", ' ', text)
tweet = tweet.lower()
tweet = tweet.split()
lem = WordNetLemmatizer()
tweet = [lem.lemmatize(word)for word in tweet
if word not in set(stopwords.words('english')) ]
tweet = ' '.join(tweet)
return tweet
print('--here goes nothing')
print(text)
print(twee... | Natural Language Processing with Disaster Tweets |
13,037,893 | prediction_df = prediction_df.sort_values('id')
prediction_df.to_csv('submission.csv', index=False )<load_from_csv> | count_vectorizer = feature_extraction.text.CountVectorizer()
train_vectors = count_vectorizer.fit_transform(train_data["text"])
test_vectors = count_vectorizer.transform(test_data["text"])
vectorizer = TfidfVectorizer()
Train = vectorizer.fit_transform(train_data['text'])
test = vectorizer.transform(test_data['text'... | Natural Language Processing with Disaster Tweets |
13,037,893 | test = pd.read_csv('.. /input/champs-scalar-coupling/test.csv')
sub1 = pd.read_csv('.. /input/keras-neural-net-and-distance-features/submission.csv')
sub2 = pd.read_csv('.. /input/keras-nn-with-multi-output/submission.csv')
display(test.head() ,sub1.head() ,sub2.head() )<create_dataframe> | clf.fit(Train, train_data["target"] ) | Natural Language Processing with Disaster Tweets |
13,037,893 | sub = pd.DataFrame(columns = ['id','scalar_coupling_constant'])
mol_types1 = ['2JHH','2JHN','2JHC','3JHH', '3JHC', '3JHN']
mol_types2 = ['1JHC', '1JHN']<concatenate> | sample_submission = pd.read_csv("/kaggle/input/nlp-getting-started/sample_submission.csv")
sample_submission["target"] = clf.predict(test)
sample_submission.head() | Natural Language Processing with Disaster Tweets |
13,037,893 | <sort_values><EOS> | sample_submission.to_csv("submission.csv", index=False ) | Natural Language Processing with Disaster Tweets |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.