In [0]:
!pip install -U -q PyDrive
from pydrive.auth import GoogleAuth
from pydrive.drive import GoogleDrive
from google.colab import auth
from oauth2client.client import GoogleCredentials
# Authenticate and create the PyDrive client.
auth.authenticate_user()
gauth = GoogleAuth()
gauth.credentials = GoogleCredentials.get_application_default()
drive = GoogleDrive(gauth)
import pandas as pd
link = 'https://drive.google.com/open?id=1ARSZM0gjAmryqyrRJff-Tkbn_CR1VxJ3'
fluff, id = link.split('=')
downloaded = drive.CreateFile({'id':id})
downloaded.GetContentFile('corpus_rev_cleaned_sample.xlsx')
link = 'https://drive.google.com/open?id=19viDsoa9J_Gn46RuVKQrpD8q4wBlejWg'
fluff, id = link.split('=')
downloaded = drive.CreateFile({'id':id})
downloaded.GetContentFile('stopwordsid.txt')
df = pd.read_excel('corpus_rev_cleaned_sample.xlsx')
In [2]:
!pip install --upgrade git+https://github.com/zalandoresearch/flair.git
Collecting git+https://github.com/zalandoresearch/flair.git
Cloning https://github.com/zalandoresearch/flair.git to /tmp/pip-req-build-ziajo1ao
Running command git clone -q https://github.com/zalandoresearch/flair.git /tmp/pip-req-build-ziajo1ao
Installing build dependencies ... done
Getting requirements to build wheel ... done
Preparing wheel metadata ... done
Collecting regex (from flair==0.4.2)
Downloading https://files.pythonhosted.org/packages/6f/4e/1b178c38c9a1a184288f72065a65ca01f3154df43c6ad898624149b8b4e0/regex-2019.06.08.tar.gz (651kB)
|████████████████████████████████| 655kB 35.5MB/s
Requirement already satisfied, skipping upgrade: gensim>=3.4.0 in /usr/local/lib/python3.6/dist-packages (from flair==0.4.2) (3.6.0)
Requirement already satisfied, skipping upgrade: tabulate in /usr/local/lib/python3.6/dist-packages (from flair==0.4.2) (0.8.3)
Collecting mpld3==0.3 (from flair==0.4.2)
Downloading https://files.pythonhosted.org/packages/91/95/a52d3a83d0a29ba0d6898f6727e9858fe7a43f6c2ce81a5fe7e05f0f4912/mpld3-0.3.tar.gz (788kB)
|████████████████████████████████| 798kB 47.6MB/s
Collecting pytorch-pretrained-bert>=0.6.1 (from flair==0.4.2)
Downloading https://files.pythonhosted.org/packages/d7/e0/c08d5553b89973d9a240605b9c12404bcf8227590de62bae27acbcfe076b/pytorch_pretrained_bert-0.6.2-py3-none-any.whl (123kB)
|████████████████████████████████| 133kB 47.1MB/s
Collecting bpemb>=0.2.9 (from flair==0.4.2)
Downloading https://files.pythonhosted.org/packages/bc/70/468a9652095b370f797ed37ff77e742b11565c6fd79eaeca5f2e50b164a7/bpemb-0.3.0-py3-none-any.whl
Requirement already satisfied, skipping upgrade: sklearn in /usr/local/lib/python3.6/dist-packages (from flair==0.4.2) (0.0)
Requirement already satisfied, skipping upgrade: urllib3<1.25,>=1.20 in /usr/local/lib/python3.6/dist-packages (from flair==0.4.2) (1.24.3)
Requirement already satisfied, skipping upgrade: tqdm>=4.26.0 in /usr/local/lib/python3.6/dist-packages (from flair==0.4.2) (4.28.1)
Requirement already satisfied, skipping upgrade: torch>=1.0.0 in /usr/local/lib/python3.6/dist-packages (from flair==0.4.2) (1.1.0)
Requirement already satisfied, skipping upgrade: matplotlib>=2.2.3 in /usr/local/lib/python3.6/dist-packages (from flair==0.4.2) (3.0.3)
Requirement already satisfied, skipping upgrade: hyperopt>=0.1.1 in /usr/local/lib/python3.6/dist-packages (from flair==0.4.2) (0.1.2)
Collecting langdetect (from flair==0.4.2)
Downloading https://files.pythonhosted.org/packages/59/59/4bc44158a767a6d66de18c4136c8aa90491d56cc951c10b74dd1e13213c9/langdetect-1.0.7.zip (998kB)
|████████████████████████████████| 1.0MB 51.1MB/s
Collecting deprecated>=1.2.4 (from flair==0.4.2)
Downloading https://files.pythonhosted.org/packages/88/0e/9d5a1a8cd7130c49334cce7b8167ceda63d6a329c8ea65b626116bc9e9e6/Deprecated-1.2.6-py2.py3-none-any.whl
Requirement already satisfied, skipping upgrade: pytest>=3.6.4 in /usr/local/lib/python3.6/dist-packages (from flair==0.4.2) (3.6.4)
Collecting segtok>=1.5.7 (from flair==0.4.2)
Downloading https://files.pythonhosted.org/packages/1d/59/6ed78856ab99d2da04084b59e7da797972baa0efecb71546b16d48e49d9b/segtok-1.5.7.tar.gz
Collecting sqlitedict>=1.6.0 (from flair==0.4.2)
Downloading https://files.pythonhosted.org/packages/0f/1c/c757b93147a219cf1e25cef7e1ad9b595b7f802159493c45ce116521caff/sqlitedict-1.6.0.tar.gz
Requirement already satisfied, skipping upgrade: six>=1.5.0 in /usr/local/lib/python3.6/dist-packages (from gensim>=3.4.0->flair==0.4.2) (1.12.0)
Requirement already satisfied, skipping upgrade: smart-open>=1.2.1 in /usr/local/lib/python3.6/dist-packages (from gensim>=3.4.0->flair==0.4.2) (1.8.4)
Requirement already satisfied, skipping upgrade: scipy>=0.18.1 in /usr/local/lib/python3.6/dist-packages (from gensim>=3.4.0->flair==0.4.2) (1.3.0)
Requirement already satisfied, skipping upgrade: numpy>=1.11.3 in /usr/local/lib/python3.6/dist-packages (from gensim>=3.4.0->flair==0.4.2) (1.16.4)
Requirement already satisfied, skipping upgrade: requests in /usr/local/lib/python3.6/dist-packages (from pytorch-pretrained-bert>=0.6.1->flair==0.4.2) (2.21.0)
Requirement already satisfied, skipping upgrade: boto3 in /usr/local/lib/python3.6/dist-packages (from pytorch-pretrained-bert>=0.6.1->flair==0.4.2) (1.9.185)
Collecting sentencepiece (from bpemb>=0.2.9->flair==0.4.2)
Downloading https://files.pythonhosted.org/packages/00/95/7f357995d5eb1131aa2092096dca14a6fc1b1d2860bd99c22a612e1d1019/sentencepiece-0.1.82-cp36-cp36m-manylinux1_x86_64.whl (1.0MB)
|████████████████████████████████| 1.0MB 50.9MB/s
Requirement already satisfied, skipping upgrade: scikit-learn in /usr/local/lib/python3.6/dist-packages (from sklearn->flair==0.4.2) (0.21.2)
Requirement already satisfied, skipping upgrade: kiwisolver>=1.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=2.2.3->flair==0.4.2) (1.1.0)
Requirement already satisfied, skipping upgrade: python-dateutil>=2.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=2.2.3->flair==0.4.2) (2.5.3)
Requirement already satisfied, skipping upgrade: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=2.2.3->flair==0.4.2) (2.4.0)
Requirement already satisfied, skipping upgrade: cycler>=0.10 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=2.2.3->flair==0.4.2) (0.10.0)
Requirement already satisfied, skipping upgrade: networkx in /usr/local/lib/python3.6/dist-packages (from hyperopt>=0.1.1->flair==0.4.2) (2.3)
Requirement already satisfied, skipping upgrade: future in /usr/local/lib/python3.6/dist-packages (from hyperopt>=0.1.1->flair==0.4.2) (0.16.0)
Requirement already satisfied, skipping upgrade: pymongo in /usr/local/lib/python3.6/dist-packages (from hyperopt>=0.1.1->flair==0.4.2) (3.8.0)
Requirement already satisfied, skipping upgrade: wrapt<2,>=1.10 in /usr/local/lib/python3.6/dist-packages (from deprecated>=1.2.4->flair==0.4.2) (1.11.2)
Requirement already satisfied, skipping upgrade: attrs>=17.4.0 in /usr/local/lib/python3.6/dist-packages (from pytest>=3.6.4->flair==0.4.2) (19.1.0)
Requirement already satisfied, skipping upgrade: atomicwrites>=1.0 in /usr/local/lib/python3.6/dist-packages (from pytest>=3.6.4->flair==0.4.2) (1.3.0)
Requirement already satisfied, skipping upgrade: setuptools in /usr/local/lib/python3.6/dist-packages (from pytest>=3.6.4->flair==0.4.2) (41.0.1)
Requirement already satisfied, skipping upgrade: pluggy<0.8,>=0.5 in /usr/local/lib/python3.6/dist-packages (from pytest>=3.6.4->flair==0.4.2) (0.7.1)
Requirement already satisfied, skipping upgrade: more-itertools>=4.0.0 in /usr/local/lib/python3.6/dist-packages (from pytest>=3.6.4->flair==0.4.2) (7.1.0)
Requirement already satisfied, skipping upgrade: py>=1.5.0 in /usr/local/lib/python3.6/dist-packages (from pytest>=3.6.4->flair==0.4.2) (1.8.0)
Requirement already satisfied, skipping upgrade: boto>=2.32 in /usr/local/lib/python3.6/dist-packages (from smart-open>=1.2.1->gensim>=3.4.0->flair==0.4.2) (2.49.0)
Requirement already satisfied, skipping upgrade: chardet<3.1.0,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests->pytorch-pretrained-bert>=0.6.1->flair==0.4.2) (3.0.4)
Requirement already satisfied, skipping upgrade: certifi>=2017.4.17 in /usr/local/lib/python3.6/dist-packages (from requests->pytorch-pretrained-bert>=0.6.1->flair==0.4.2) (2019.6.16)
Requirement already satisfied, skipping upgrade: idna<2.9,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests->pytorch-pretrained-bert>=0.6.1->flair==0.4.2) (2.8)
Requirement already satisfied, skipping upgrade: jmespath<1.0.0,>=0.7.1 in /usr/local/lib/python3.6/dist-packages (from boto3->pytorch-pretrained-bert>=0.6.1->flair==0.4.2) (0.9.4)
Requirement already satisfied, skipping upgrade: s3transfer<0.3.0,>=0.2.0 in /usr/local/lib/python3.6/dist-packages (from boto3->pytorch-pretrained-bert>=0.6.1->flair==0.4.2) (0.2.1)
Requirement already satisfied, skipping upgrade: botocore<1.13.0,>=1.12.185 in /usr/local/lib/python3.6/dist-packages (from boto3->pytorch-pretrained-bert>=0.6.1->flair==0.4.2) (1.12.185)
Requirement already satisfied, skipping upgrade: joblib>=0.11 in /usr/local/lib/python3.6/dist-packages (from scikit-learn->sklearn->flair==0.4.2) (0.13.2)
Requirement already satisfied, skipping upgrade: decorator>=4.3.0 in /usr/local/lib/python3.6/dist-packages (from networkx->hyperopt>=0.1.1->flair==0.4.2) (4.4.0)
Requirement already satisfied, skipping upgrade: docutils>=0.10 in /usr/local/lib/python3.6/dist-packages (from botocore<1.13.0,>=1.12.185->boto3->pytorch-pretrained-bert>=0.6.1->flair==0.4.2) (0.14)
Building wheels for collected packages: flair
Building wheel for flair (PEP 517) ... done
Stored in directory: /tmp/pip-ephem-wheel-cache-s7lkfavx/wheels/6a/78/0f/399330241d3bc69458cc4fe320dcdfbf818f9887803f0294e7
Successfully built flair
Building wheels for collected packages: regex, mpld3, langdetect, segtok, sqlitedict
Building wheel for regex (setup.py) ... done
Stored in directory: /root/.cache/pip/wheels/35/e4/80/abf3b33ba89cf65cd262af8a22a5a999cc28fbfabea6b38473
Building wheel for mpld3 (setup.py) ... done
Stored in directory: /root/.cache/pip/wheels/c0/47/fb/8a64f89aecfe0059830479308ad42d62e898a3e3cefdf6ba28
Building wheel for langdetect (setup.py) ... done
Stored in directory: /root/.cache/pip/wheels/ec/0c/a9/1647275e7ef5014e7b83ff30105180e332867d65e7617ddafe
Building wheel for segtok (setup.py) ... done
Stored in directory: /root/.cache/pip/wheels/15/ee/a8/6112173f1386d33eebedb3f73429cfa41a4c3084556bcee254
Building wheel for sqlitedict (setup.py) ... done
Stored in directory: /root/.cache/pip/wheels/bd/57/d3/907c3ee02d35e66f674ad0106e61f06eeeb98f6ee66a6cc3fe
Successfully built regex mpld3 langdetect segtok sqlitedict
Installing collected packages: regex, mpld3, pytorch-pretrained-bert, sentencepiece, bpemb, langdetect, deprecated, segtok, sqlitedict, flair
Successfully installed bpemb-0.3.0 deprecated-1.2.6 flair-0.4.2 langdetect-1.0.7 mpld3-0.3 pytorch-pretrained-bert-0.6.2 regex-2019.6.8 segtok-1.5.7 sentencepiece-0.1.82 sqlitedict-1.6.0
In [17]:
import pandas as pd
import numpy as np
import re
f = open("stopwordsid.txt", 'r')
stop_words = f.readlines()
f.close()
def removestopWords(s):
s = ' '.join(word for word in s.split() if word not in stop_words)
return s
df = pd.read_excel('corpus_rev_cleaned_sample.xlsx')
df = df.drop(['publish_date','doc_id','sentence_num','sentence_prev','sentence_next'],axis=1)
df['sentence'] = df['sentence'].apply(lambda x: " ".join(x.lower() for x in str(x).split()))
df['sentence'] = df['sentence'].apply(lambda x : " ".join(re.findall('[\w]+',x)))
df['sentence'] = df['sentence'].apply(lambda x: removestopWords(x))
df = df.dropna()
df = df.reset_index(drop=True)
df['credibility'] = np.where(((df['formulasi']=='-')&(df['efektivitas']=='-')&(df['koordinasi']=='-')&(df['stance komunikasi']=='-')),0,1)
df.head()
Out[17]:
sentence
formulasi
efektivitas
koordinasi
stance komunikasi
credibility
0
kalau bisa menekan suku bunga kredit lebih ren...
-
-
-
-
0
1
perbankan masih menggunakan suku bunga lama se...
-
1
-
-
1
2
penurunan bi rate juga bisa membuat pergerakan...
-
1
-
-
1
3
kebijakan suku bunga ditempuh oleh bi merupaka...
1
-
-
-
1
4
penyesuaian bi rate akan memperkuat operasi mo...
1
1
-
-
1
In [0]:
link = 'https://drive.google.com/open?id=1o3SbepNJm3cCNkkF_7ROuwuSMyhpheL0'
fluff, id = link.split('=')
downloaded = drive.CreateFile({'id':id})
downloaded.GetContentFile('test_sample.csv')
In [25]:
dftest = pd.read_csv('test_sample.csv')
dftest['credibility'] = np.where(((dftest['formulasi']=='-')&(dftest['efektivitas']=='-')&(dftest['koordinasi']=='-')&(dftest['stance komunikasi']=='-')),0,1)
print(len(df))
len(dftest)
3674
Out[25]:
502
In [30]:
import os
import random
os.remove("train.txt")
os.remove("test.txt")
os.remove("dev.txt")
import math
listrand = list(range(100))
random.shuffle(listrand)
listranddev = []
listrandtest = []
for i in range(math.floor(len(listrand)*0.2)):
listranddev.append(listrand.pop())
file1 = open("train.txt","a", encoding="utf-8")
for i in listrand:
file1.writelines(['__label__',str(df['credibility'].iloc[i]),' ',str(df['sentence'].iloc[i]),'\n'])
file1.close()
file2 = open("dev.txt","a", encoding="utf-8")
for i in listranddev:
file2.writelines(['__label__',str(df['credibility'].iloc[i]),' ',str(df['sentence'].iloc[i]),'\n'])
file2.close()
file3 = open("test.txt","a", encoding="utf-8")
for i in range(len(dftest)):
file3.writelines(['__label__',str(dftest['credibility'].iloc[i]),' ',str(dftest['sentence'].iloc[i]),'\n'])
file3.close()
[27, 12, 26, 45, 92, 22, 80, 66, 59, 95]
In [50]:
from flair.data_fetcher import NLPTaskDataFetcher
from flair.embeddings import WordEmbeddings, FlairEmbeddings, DocumentRNNEmbeddings
from flair.models import TextClassifier
from flair.trainers import ModelTrainer
from pathlib import Path
corpus = NLPTaskDataFetcher.load_classification_corpus(Path('./'), test_file='test.txt', dev_file='dev.txt', train_file='train.txt')
word_embeddings = [WordEmbeddings('glove'), FlairEmbeddings('id-forward'), FlairEmbeddings('id-backward')]
document_embeddings = DocumentRNNEmbeddings(word_embeddings, hidden_size=256, bidirectional=True,rnn_type='lstm',rnn_layers=1)
classifier = TextClassifier(document_embeddings, label_dictionary=corpus.make_label_dictionary(), multi_label=False)
trainer = ModelTrainer(classifier, corpus)
trainer.train('./', max_epochs=30,checkpoint=True)
!mv best-model.pt best-cred-model.pt
2019-07-17 13:56:46,627 Reading data from .
2019-07-17 13:56:46,630 Train: train.txt
2019-07-17 13:56:46,631 Dev: dev.txt
2019-07-17 13:56:46,635 Test: test.txt
/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:7: DeprecationWarning: Call to deprecated function (or staticmethod) load_classification_corpus. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
import sys
/usr/local/lib/python3.6/dist-packages/flair/data_fetcher.py:447: DeprecationWarning: Call to deprecated function (or staticmethod) read_text_classification_file. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
max_tokens_per_doc=max_tokens_per_doc,
/usr/local/lib/python3.6/dist-packages/flair/data_fetcher.py:454: DeprecationWarning: Call to deprecated function (or staticmethod) read_text_classification_file. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
max_tokens_per_doc=max_tokens_per_doc,
/usr/local/lib/python3.6/dist-packages/flair/data_fetcher.py:463: DeprecationWarning: Call to deprecated function (or staticmethod) read_text_classification_file. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
max_tokens_per_doc=max_tokens_per_doc,
/usr/local/lib/python3.6/dist-packages/smart_open/smart_open_lib.py:398: UserWarning: This function is deprecated, use smart_open.open instead. See the migration notes for details: https://github.com/RaRe-Technologies/smart_open/blob/master/README.rst#migrating-to-the-new-open-function
'See the migration notes for details: %s' % _MIGRATION_NOTES_URL
2019-07-17 13:56:50,251 Computing label dictionary. Progress:
100%|██████████| 80/80 [00:00<00:00, 73632.72it/s]
2019-07-17 13:56:50,259 [b'1', b'0']
2019-07-17 13:56:50,485 ----------------------------------------------------------------------------------------------------
2019-07-17 13:56:50,487 Model: "TextClassifier(
(document_embeddings): DocumentRNNEmbeddings(
(embeddings): StackedEmbeddings(
(list_embedding_0): WordEmbeddings('glove')
(list_embedding_1): FlairEmbeddings(
(lm): LanguageModel(
(drop): Dropout(p=0.1)
(encoder): Embedding(8823, 100)
(rnn): LSTM(100, 2048)
(decoder): Linear(in_features=2048, out_features=8823, bias=True)
)
)
(list_embedding_2): FlairEmbeddings(
(lm): LanguageModel(
(drop): Dropout(p=0.1)
(encoder): Embedding(8823, 100)
(rnn): LSTM(100, 2048)
(decoder): Linear(in_features=2048, out_features=8823, bias=True)
)
)
)
(word_reprojection_map): Linear(in_features=4196, out_features=4196, bias=True)
(rnn): GRU(4196, 256, bidirectional=True)
(dropout): Dropout(p=0.5)
)
(decoder): Linear(in_features=1024, out_features=2, bias=True)
(loss_function): CrossEntropyLoss()
)"
2019-07-17 13:56:50,490 ----------------------------------------------------------------------------------------------------
2019-07-17 13:56:50,492 Corpus: "Corpus: 80 train + 20 dev + 502 test sentences"
2019-07-17 13:56:50,494 ----------------------------------------------------------------------------------------------------
2019-07-17 13:56:50,495 Parameters:
2019-07-17 13:56:50,496 - learning_rate: "0.1"
2019-07-17 13:56:50,498 - mini_batch_size: "32"
2019-07-17 13:56:50,499 - patience: "3"
2019-07-17 13:56:50,500 - anneal_factor: "0.5"
2019-07-17 13:56:50,501 - max_epochs: "30"
2019-07-17 13:56:50,503 - shuffle: "True"
2019-07-17 13:56:50,504 - train_with_dev: "False"
2019-07-17 13:56:50,505 ----------------------------------------------------------------------------------------------------
2019-07-17 13:56:50,507 Model training base path: "."
2019-07-17 13:56:50,508 ----------------------------------------------------------------------------------------------------
2019-07-17 13:56:50,509 Device: cuda:0
2019-07-17 13:56:50,510 ----------------------------------------------------------------------------------------------------
2019-07-17 13:56:50,512 Embedding storage mode: cpu
2019-07-17 13:56:50,514 ----------------------------------------------------------------------------------------------------
2019-07-17 13:56:51,138 epoch 1 - iter 0/3 - loss 0.87486637
2019-07-17 13:56:51,679 epoch 1 - iter 1/3 - loss 1.03721493
2019-07-17 13:56:52,000 epoch 1 - iter 2/3 - loss 1.09956042
2019-07-17 13:56:52,017 ----------------------------------------------------------------------------------------------------
2019-07-17 13:56:52,018 EPOCH 1 done: loss 1.0996 - lr 0.1000
2019-07-17 13:56:52,367 DEV : loss 0.9419633746147156 - score 0.55
2019-07-17 13:56:52,405 BAD EPOCHS (no improvement): 0
2019-07-17 13:57:04,492 ----------------------------------------------------------------------------------------------------
2019-07-17 13:57:04,596 epoch 2 - iter 0/3 - loss 0.64649445
2019-07-17 13:57:04,754 epoch 2 - iter 1/3 - loss 0.76610887
2019-07-17 13:57:04,821 epoch 2 - iter 2/3 - loss 0.98244977
2019-07-17 13:57:04,840 ----------------------------------------------------------------------------------------------------
2019-07-17 13:57:04,841 EPOCH 2 done: loss 0.9824 - lr 0.1000
2019-07-17 13:57:04,877 DEV : loss 0.781875729560852 - score 0.55
2019-07-17 13:57:04,882 BAD EPOCHS (no improvement): 1
2019-07-17 13:57:16,329 ----------------------------------------------------------------------------------------------------
2019-07-17 13:57:16,417 epoch 3 - iter 0/3 - loss 0.41489190
2019-07-17 13:57:16,533 epoch 3 - iter 1/3 - loss 0.37624091
2019-07-17 13:57:16,623 epoch 3 - iter 2/3 - loss 0.59396513
2019-07-17 13:57:16,641 ----------------------------------------------------------------------------------------------------
2019-07-17 13:57:16,642 EPOCH 3 done: loss 0.5940 - lr 0.1000
2019-07-17 13:57:16,678 DEV : loss 1.1844289302825928 - score 0.45
2019-07-17 13:57:16,682 BAD EPOCHS (no improvement): 2
2019-07-17 13:57:22,390 ----------------------------------------------------------------------------------------------------
2019-07-17 13:57:22,483 epoch 4 - iter 0/3 - loss 0.78919774
2019-07-17 13:57:22,602 epoch 4 - iter 1/3 - loss 0.65593708
2019-07-17 13:57:22,672 epoch 4 - iter 2/3 - loss 0.55941910
2019-07-17 13:57:22,690 ----------------------------------------------------------------------------------------------------
2019-07-17 13:57:22,691 EPOCH 4 done: loss 0.5594 - lr 0.1000
2019-07-17 13:57:22,724 DEV : loss 0.7642226219177246 - score 0.65
2019-07-17 13:57:22,728 BAD EPOCHS (no improvement): 0
2019-07-17 13:57:34,373 ----------------------------------------------------------------------------------------------------
2019-07-17 13:57:34,469 epoch 5 - iter 0/3 - loss 0.23155423
2019-07-17 13:57:34,624 epoch 5 - iter 1/3 - loss 0.27474933
2019-07-17 13:57:34,688 epoch 5 - iter 2/3 - loss 0.27912891
2019-07-17 13:57:34,706 ----------------------------------------------------------------------------------------------------
2019-07-17 13:57:34,707 EPOCH 5 done: loss 0.2791 - lr 0.1000
2019-07-17 13:57:34,742 DEV : loss 0.7938563227653503 - score 0.7
2019-07-17 13:57:34,746 BAD EPOCHS (no improvement): 0
2019-07-17 13:57:46,178 ----------------------------------------------------------------------------------------------------
2019-07-17 13:57:46,480 epoch 6 - iter 0/3 - loss 0.24905553
2019-07-17 13:57:46,580 epoch 6 - iter 1/3 - loss 0.23486228
2019-07-17 13:57:46,652 epoch 6 - iter 2/3 - loss 0.22392610
2019-07-17 13:57:46,669 ----------------------------------------------------------------------------------------------------
2019-07-17 13:57:46,670 EPOCH 6 done: loss 0.2239 - lr 0.1000
2019-07-17 13:57:46,711 DEV : loss 0.8911200761795044 - score 0.55
2019-07-17 13:57:46,715 BAD EPOCHS (no improvement): 1
2019-07-17 13:57:52,488 ----------------------------------------------------------------------------------------------------
2019-07-17 13:57:52,588 epoch 7 - iter 0/3 - loss 0.33576858
2019-07-17 13:57:52,699 epoch 7 - iter 1/3 - loss 0.36381039
2019-07-17 13:57:52,767 epoch 7 - iter 2/3 - loss 0.31264396
2019-07-17 13:57:52,785 ----------------------------------------------------------------------------------------------------
2019-07-17 13:57:52,786 EPOCH 7 done: loss 0.3126 - lr 0.1000
2019-07-17 13:57:52,819 DEV : loss 0.8823937177658081 - score 0.65
2019-07-17 13:57:52,823 BAD EPOCHS (no improvement): 2
2019-07-17 13:57:58,512 ----------------------------------------------------------------------------------------------------
2019-07-17 13:57:58,611 epoch 8 - iter 0/3 - loss 0.16058338
2019-07-17 13:57:58,743 epoch 8 - iter 1/3 - loss 0.16332692
2019-07-17 13:57:58,809 epoch 8 - iter 2/3 - loss 0.22525620
2019-07-17 13:57:58,827 ----------------------------------------------------------------------------------------------------
2019-07-17 13:57:58,828 EPOCH 8 done: loss 0.2253 - lr 0.1000
2019-07-17 13:57:58,861 DEV : loss 0.9370827674865723 - score 0.6
2019-07-17 13:57:58,864 BAD EPOCHS (no improvement): 3
2019-07-17 13:58:04,549 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:04,657 epoch 9 - iter 0/3 - loss 0.19470981
2019-07-17 13:58:04,763 epoch 9 - iter 1/3 - loss 0.16204910
2019-07-17 13:58:04,831 epoch 9 - iter 2/3 - loss 0.15333315
2019-07-17 13:58:04,849 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:04,850 EPOCH 9 done: loss 0.1533 - lr 0.1000
2019-07-17 13:58:04,884 DEV : loss 0.9461809992790222 - score 0.55
Epoch 8: reducing learning rate of group 0 to 5.0000e-02.
2019-07-17 13:58:04,890 BAD EPOCHS (no improvement): 4
2019-07-17 13:58:10,564 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:10,663 epoch 10 - iter 0/3 - loss 0.11734888
2019-07-17 13:58:10,793 epoch 10 - iter 1/3 - loss 0.11035379
2019-07-17 13:58:10,860 epoch 10 - iter 2/3 - loss 0.10750741
2019-07-17 13:58:10,880 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:10,881 EPOCH 10 done: loss 0.1075 - lr 0.0500
2019-07-17 13:58:10,914 DEV : loss 0.8466167449951172 - score 0.65
2019-07-17 13:58:10,918 BAD EPOCHS (no improvement): 1
2019-07-17 13:58:16,592 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:16,702 epoch 11 - iter 0/3 - loss 0.07047906
2019-07-17 13:58:16,827 epoch 11 - iter 1/3 - loss 0.07457037
2019-07-17 13:58:16,889 epoch 11 - iter 2/3 - loss 0.08159962
2019-07-17 13:58:16,906 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:16,907 EPOCH 11 done: loss 0.0816 - lr 0.0500
2019-07-17 13:58:16,942 DEV : loss 0.8569294214248657 - score 0.65
2019-07-17 13:58:16,946 BAD EPOCHS (no improvement): 2
2019-07-17 13:58:22,616 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:22,707 epoch 12 - iter 0/3 - loss 0.09211741
2019-07-17 13:58:22,832 epoch 12 - iter 1/3 - loss 0.08469783
2019-07-17 13:58:22,898 epoch 12 - iter 2/3 - loss 0.07519636
2019-07-17 13:58:22,916 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:22,917 EPOCH 12 done: loss 0.0752 - lr 0.0500
2019-07-17 13:58:22,951 DEV : loss 0.876681923866272 - score 0.65
2019-07-17 13:58:22,956 BAD EPOCHS (no improvement): 3
2019-07-17 13:58:28,674 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:28,777 epoch 13 - iter 0/3 - loss 0.07967311
2019-07-17 13:58:28,916 epoch 13 - iter 1/3 - loss 0.07925714
2019-07-17 13:58:28,974 epoch 13 - iter 2/3 - loss 0.07909881
2019-07-17 13:58:28,993 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:28,994 EPOCH 13 done: loss 0.0791 - lr 0.0500
2019-07-17 13:58:29,027 DEV : loss 0.8545287847518921 - score 0.7
Epoch 12: reducing learning rate of group 0 to 2.5000e-02.
2019-07-17 13:58:29,032 BAD EPOCHS (no improvement): 4
2019-07-17 13:58:40,957 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:41,187 epoch 14 - iter 0/3 - loss 0.06494763
2019-07-17 13:58:41,288 epoch 14 - iter 1/3 - loss 0.08248516
2019-07-17 13:58:41,350 epoch 14 - iter 2/3 - loss 0.08677889
2019-07-17 13:58:41,368 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:41,368 EPOCH 14 done: loss 0.0868 - lr 0.0250
2019-07-17 13:58:41,401 DEV : loss 0.8664800524711609 - score 0.65
2019-07-17 13:58:41,405 BAD EPOCHS (no improvement): 1
2019-07-17 13:58:47,168 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:47,265 epoch 15 - iter 0/3 - loss 0.06638304
2019-07-17 13:58:47,373 epoch 15 - iter 1/3 - loss 0.07129348
2019-07-17 13:58:47,445 epoch 15 - iter 2/3 - loss 0.08240320
2019-07-17 13:58:47,463 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:47,464 EPOCH 15 done: loss 0.0824 - lr 0.0250
2019-07-17 13:58:47,496 DEV : loss 0.8889740109443665 - score 0.65
2019-07-17 13:58:47,500 BAD EPOCHS (no improvement): 2
2019-07-17 13:58:53,216 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:53,308 epoch 16 - iter 0/3 - loss 0.08196284
2019-07-17 13:58:53,427 epoch 16 - iter 1/3 - loss 0.07381566
2019-07-17 13:58:53,501 epoch 16 - iter 2/3 - loss 0.07220588
2019-07-17 13:58:53,520 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:53,521 EPOCH 16 done: loss 0.0722 - lr 0.0250
2019-07-17 13:58:53,556 DEV : loss 0.8935390710830688 - score 0.65
2019-07-17 13:58:53,561 BAD EPOCHS (no improvement): 3
2019-07-17 13:58:59,426 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:59,525 epoch 17 - iter 0/3 - loss 0.05390355
2019-07-17 13:58:59,667 epoch 17 - iter 1/3 - loss 0.05964504
2019-07-17 13:58:59,728 epoch 17 - iter 2/3 - loss 0.06481090
2019-07-17 13:58:59,745 ----------------------------------------------------------------------------------------------------
2019-07-17 13:58:59,746 EPOCH 17 done: loss 0.0648 - lr 0.0250
2019-07-17 13:58:59,781 DEV : loss 0.8994463682174683 - score 0.65
Epoch 16: reducing learning rate of group 0 to 1.2500e-02.
2019-07-17 13:58:59,786 BAD EPOCHS (no improvement): 4
2019-07-17 13:59:05,472 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:05,570 epoch 18 - iter 0/3 - loss 0.05169375
2019-07-17 13:59:05,685 epoch 18 - iter 1/3 - loss 0.05780241
2019-07-17 13:59:05,761 epoch 18 - iter 2/3 - loss 0.05412370
2019-07-17 13:59:05,778 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:05,779 EPOCH 18 done: loss 0.0541 - lr 0.0125
2019-07-17 13:59:05,815 DEV : loss 0.9014705419540405 - score 0.65
2019-07-17 13:59:05,818 BAD EPOCHS (no improvement): 1
2019-07-17 13:59:11,567 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:11,672 epoch 19 - iter 0/3 - loss 0.07720007
2019-07-17 13:59:11,787 epoch 19 - iter 1/3 - loss 0.07120714
2019-07-17 13:59:11,853 epoch 19 - iter 2/3 - loss 0.07041672
2019-07-17 13:59:11,871 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:11,872 EPOCH 19 done: loss 0.0704 - lr 0.0125
2019-07-17 13:59:11,907 DEV : loss 0.9066303968429565 - score 0.65
2019-07-17 13:59:11,911 BAD EPOCHS (no improvement): 2
2019-07-17 13:59:17,618 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:17,726 epoch 20 - iter 0/3 - loss 0.08289422
2019-07-17 13:59:17,839 epoch 20 - iter 1/3 - loss 0.06373616
2019-07-17 13:59:17,921 epoch 20 - iter 2/3 - loss 0.06625385
2019-07-17 13:59:17,940 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:17,941 EPOCH 20 done: loss 0.0663 - lr 0.0125
2019-07-17 13:59:17,975 DEV : loss 0.9165571331977844 - score 0.65
2019-07-17 13:59:17,980 BAD EPOCHS (no improvement): 3
2019-07-17 13:59:23,694 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:23,790 epoch 21 - iter 0/3 - loss 0.03951208
2019-07-17 13:59:23,912 epoch 21 - iter 1/3 - loss 0.05145373
2019-07-17 13:59:23,972 epoch 21 - iter 2/3 - loss 0.05428546
2019-07-17 13:59:23,989 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:23,990 EPOCH 21 done: loss 0.0543 - lr 0.0125
2019-07-17 13:59:24,024 DEV : loss 0.9184209704399109 - score 0.65
Epoch 20: reducing learning rate of group 0 to 6.2500e-03.
2019-07-17 13:59:24,028 BAD EPOCHS (no improvement): 4
2019-07-17 13:59:29,943 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:30,042 epoch 22 - iter 0/3 - loss 0.04506227
2019-07-17 13:59:30,168 epoch 22 - iter 1/3 - loss 0.05107411
2019-07-17 13:59:30,223 epoch 22 - iter 2/3 - loss 0.05973764
2019-07-17 13:59:30,241 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:30,242 EPOCH 22 done: loss 0.0597 - lr 0.0063
2019-07-17 13:59:30,273 DEV : loss 0.9208900332450867 - score 0.65
2019-07-17 13:59:30,277 BAD EPOCHS (no improvement): 1
2019-07-17 13:59:36,066 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:36,170 epoch 23 - iter 0/3 - loss 0.04833972
2019-07-17 13:59:36,286 epoch 23 - iter 1/3 - loss 0.05488302
2019-07-17 13:59:36,358 epoch 23 - iter 2/3 - loss 0.05602942
2019-07-17 13:59:36,375 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:36,376 EPOCH 23 done: loss 0.0560 - lr 0.0063
2019-07-17 13:59:36,411 DEV : loss 0.9235916137695312 - score 0.65
2019-07-17 13:59:36,415 BAD EPOCHS (no improvement): 2
2019-07-17 13:59:42,246 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:42,342 epoch 24 - iter 0/3 - loss 0.05683773
2019-07-17 13:59:42,454 epoch 24 - iter 1/3 - loss 0.05273701
2019-07-17 13:59:42,541 epoch 24 - iter 2/3 - loss 0.05377713
2019-07-17 13:59:42,559 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:42,559 EPOCH 24 done: loss 0.0538 - lr 0.0063
2019-07-17 13:59:42,597 DEV : loss 0.9255558252334595 - score 0.65
2019-07-17 13:59:42,601 BAD EPOCHS (no improvement): 3
2019-07-17 13:59:48,307 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:48,413 epoch 25 - iter 0/3 - loss 0.06859999
2019-07-17 13:59:48,538 epoch 25 - iter 1/3 - loss 0.06768823
2019-07-17 13:59:48,605 epoch 25 - iter 2/3 - loss 0.05368160
2019-07-17 13:59:48,623 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:48,624 EPOCH 25 done: loss 0.0537 - lr 0.0063
2019-07-17 13:59:48,658 DEV : loss 0.921922504901886 - score 0.65
Epoch 24: reducing learning rate of group 0 to 3.1250e-03.
2019-07-17 13:59:48,662 BAD EPOCHS (no improvement): 4
2019-07-17 13:59:54,404 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:54,508 epoch 26 - iter 0/3 - loss 0.07323258
2019-07-17 13:59:54,632 epoch 26 - iter 1/3 - loss 0.05841864
2019-07-17 13:59:54,697 epoch 26 - iter 2/3 - loss 0.06093185
2019-07-17 13:59:54,717 ----------------------------------------------------------------------------------------------------
2019-07-17 13:59:54,718 EPOCH 26 done: loss 0.0609 - lr 0.0031
2019-07-17 13:59:54,752 DEV : loss 0.9239110946655273 - score 0.65
2019-07-17 13:59:54,756 BAD EPOCHS (no improvement): 1
2019-07-17 14:00:00,414 ----------------------------------------------------------------------------------------------------
2019-07-17 14:00:00,518 epoch 27 - iter 0/3 - loss 0.05320657
2019-07-17 14:00:00,627 epoch 27 - iter 1/3 - loss 0.05368141
2019-07-17 14:00:00,700 epoch 27 - iter 2/3 - loss 0.05869312
2019-07-17 14:00:00,717 ----------------------------------------------------------------------------------------------------
2019-07-17 14:00:00,718 EPOCH 27 done: loss 0.0587 - lr 0.0031
2019-07-17 14:00:00,750 DEV : loss 0.9228469133377075 - score 0.65
2019-07-17 14:00:00,754 BAD EPOCHS (no improvement): 2
2019-07-17 14:00:06,643 ----------------------------------------------------------------------------------------------------
2019-07-17 14:00:06,915 epoch 28 - iter 0/3 - loss 0.06497187
2019-07-17 14:00:07,022 epoch 28 - iter 1/3 - loss 0.06028974
2019-07-17 14:00:07,093 epoch 28 - iter 2/3 - loss 0.05676010
2019-07-17 14:00:07,111 ----------------------------------------------------------------------------------------------------
2019-07-17 14:00:07,111 EPOCH 28 done: loss 0.0568 - lr 0.0031
2019-07-17 14:00:07,145 DEV : loss 0.9234123229980469 - score 0.65
2019-07-17 14:00:07,149 BAD EPOCHS (no improvement): 3
2019-07-17 14:00:12,795 ----------------------------------------------------------------------------------------------------
2019-07-17 14:00:12,881 epoch 29 - iter 0/3 - loss 0.03929017
2019-07-17 14:00:13,003 epoch 29 - iter 1/3 - loss 0.04207543
2019-07-17 14:00:13,091 epoch 29 - iter 2/3 - loss 0.04506969
2019-07-17 14:00:13,109 ----------------------------------------------------------------------------------------------------
2019-07-17 14:00:13,110 EPOCH 29 done: loss 0.0451 - lr 0.0031
2019-07-17 14:00:13,147 DEV : loss 0.9244791269302368 - score 0.65
Epoch 28: reducing learning rate of group 0 to 1.5625e-03.
2019-07-17 14:00:13,151 BAD EPOCHS (no improvement): 4
2019-07-17 14:00:19,443 ----------------------------------------------------------------------------------------------------
2019-07-17 14:00:19,555 epoch 30 - iter 0/3 - loss 0.07640116
2019-07-17 14:00:19,674 epoch 30 - iter 1/3 - loss 0.06439571
2019-07-17 14:00:19,753 epoch 30 - iter 2/3 - loss 0.06208072
2019-07-17 14:00:19,776 ----------------------------------------------------------------------------------------------------
2019-07-17 14:00:19,778 EPOCH 30 done: loss 0.0621 - lr 0.0016
2019-07-17 14:00:19,821 DEV : loss 0.9255158305168152 - score 0.65
2019-07-17 14:00:19,826 BAD EPOCHS (no improvement): 1
2019-07-17 14:00:31,622 ----------------------------------------------------------------------------------------------------
2019-07-17 14:00:31,763 Testing using best model ...
2019-07-17 14:00:31,765 loading file best-model.pt
2019-07-17 14:00:45,893 0.5777 0.5777 0.5777
2019-07-17 14:00:45,894
MICRO_AVG: acc 0.4062 - f1-score 0.5777
MACRO_AVG: acc 0.3966 - f1-score 0.56365
0 tp: 190 - fp: 160 - fn: 52 - tn: 100 - precision: 0.5429 - recall: 0.7851 - accuracy: 0.4726 - f1-score: 0.6419
1 tp: 100 - fp: 52 - fn: 160 - tn: 190 - precision: 0.6579 - recall: 0.3846 - accuracy: 0.3205 - f1-score: 0.4854
2019-07-17 14:00:45,895 ----------------------------------------------------------------------------------------------------
Out[50]:
{'dev_loss_history': [tensor(0.9420, device='cuda:0'),
tensor(0.7819, device='cuda:0'),
tensor(1.1844, device='cuda:0'),
tensor(0.7642, device='cuda:0'),
tensor(0.7939, device='cuda:0'),
tensor(0.8911, device='cuda:0'),
tensor(0.8824, device='cuda:0'),
tensor(0.9371, device='cuda:0'),
tensor(0.9462, device='cuda:0'),
tensor(0.8466, device='cuda:0'),
tensor(0.8569, device='cuda:0'),
tensor(0.8767, device='cuda:0'),
tensor(0.8545, device='cuda:0'),
tensor(0.8665, device='cuda:0'),
tensor(0.8890, device='cuda:0'),
tensor(0.8935, device='cuda:0'),
tensor(0.8994, device='cuda:0'),
tensor(0.9015, device='cuda:0'),
tensor(0.9066, device='cuda:0'),
tensor(0.9166, device='cuda:0'),
tensor(0.9184, device='cuda:0'),
tensor(0.9209, device='cuda:0'),
tensor(0.9236, device='cuda:0'),
tensor(0.9256, device='cuda:0'),
tensor(0.9219, device='cuda:0'),
tensor(0.9239, device='cuda:0'),
tensor(0.9228, device='cuda:0'),
tensor(0.9234, device='cuda:0'),
tensor(0.9245, device='cuda:0'),
tensor(0.9255, device='cuda:0')],
'dev_score_history': [0.55,
0.55,
0.45,
0.65,
0.7,
0.55,
0.65,
0.6,
0.55,
0.65,
0.65,
0.65,
0.7,
0.65,
0.65,
0.65,
0.65,
0.65,
0.65,
0.65,
0.65,
0.65,
0.65,
0.65,
0.65,
0.65,
0.65,
0.65,
0.65,
0.65],
'test_score': 0.5777,
'train_loss_history': [1.0995604197184246,
0.9824497699737549,
0.5939651330312093,
0.5594190955162048,
0.27912891407807666,
0.2239260971546173,
0.31264396011829376,
0.22525620460510254,
0.15333314736684164,
0.10750740766525269,
0.08159961799780528,
0.07519636179010074,
0.07909881323575974,
0.08677889158328374,
0.08240320285161336,
0.07220587879419327,
0.06481090063850085,
0.054123700906833015,
0.07041671872138977,
0.06625384837388992,
0.05428546170393626,
0.05973763515551885,
0.05602941537896792,
0.053777127216259636,
0.053681597113609314,
0.06093184898296992,
0.05869312211871147,
0.056760103752215706,
0.045069692035516105,
0.062080717335144676]}
In [52]:
!ls
adc.json devkomm.txt testeff.txt traineff.txt
best-cred-model.pt devkoor.txt testform.txt trainform.txt
checkpoint.pt dev.txt testkomm.txt training.log
corpus_rev_cleaned_sample.xlsx final-model.pt testkoor.txt trainkomm.txt
corpus_rev_cleaned.xlsx loss.tsv test_sample.csv trainkoor.txt
deveff.txt sample_data test.tsv train.txt
devform.txt stopwordsid.txt test.txt weights.txt
In [32]:
dfcred = df.loc[df['credibility']==1]
dfcred.head()
dfcred['formulasi'] = dfcred['formulasi'].replace({'1':1,'X':0,-1:2,'-1':2,0:2,'0':2,'-':0})
dfcred['efektivitas'] = dfcred['efektivitas'].replace({'1':1,'X':0,-1:2,'-1':2,0:2,'0':2,'-':0})
dfcred['koordinasi'] = dfcred['koordinasi'].replace({'1':1,'X':0,-1:2,'-1':2,0:2,'0':2,'-':0})
dfcred['stance komunikasi'] = dfcred['stance komunikasi'].replace({'1':1,'X':0,-1:2,'-1':2,0:3,'0':3,'-':0})
set(dfcred['formulasi'])
/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:3: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead
See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy
This is separate from the ipykernel package so we can avoid doing imports until
/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:4: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead
See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy
after removing the cwd from sys.path.
/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:5: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead
See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy
"""
/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:6: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead
See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy
Out[32]:
{0, 1, 2}
In [39]:
dftestcred = dftest.loc[dftest['credibility']==1]
dftestcred = dftestcred.dropna()
dftestcred['formulasi'] = dftestcred['formulasi'].replace({'1':1,'X':0,-1:2,'-1':2,0:2,'0':2,'-':0})
dftestcred['efektivitas'] = dftestcred['efektivitas'].replace({'1':1,'X':0,-1:2,'-1':2,0:2,'0':2,'-':0})
dftestcred['koordinasi'] = dftestcred['koordinasi'].replace({'1':1,'X':0,-1:2,'-1':2,0:2,'0':2,'-':0})
dftestcred['stance komunikasi'] = dftestcred['stance komunikasi'].replace({'1':1,'X':0,-1:2,'-1':2,0:3,'0':3,'-':0})
print(len(dftestcred))
257
In [0]:
import os
import math
import random
# os.remove("trainform.txt")
# os.remove("testform.txt")
# os.remove("devform.txt")
listrand = list(range(100))
random.shuffle(listrand)
listranddev = []
for i in range(math.floor(len(listrand)*0.2)):
listranddev.append(listrand.pop())
file1 = open("trainform.txt","a", encoding="utf-8")
for i in listrand:
file1.writelines(['__label__',str(dfcred['formulasi'].iloc[i]),' ',str(dfcred['sentence'].iloc[i]),'\n'])
file1.close()
file2 = open("devform.txt","a", encoding="utf-8")
for i in listranddev:
file2.writelines(['__label__',str(dfcred['formulasi'].iloc[i]),' ',str(dfcred['sentence'].iloc[i]),'\n'])
file2.close()
file3 = open("testform.txt","a", encoding="utf-8")
for i in range(len(dftestcred)):
file3.writelines(['__label__',str(dftestcred['formulasi'].iloc[i]),' ',str(dftestcred['sentence'].iloc[i]),'\n'])
file3.close()
In [53]:
from flair.data_fetcher import NLPTaskDataFetcher
from flair.embeddings import WordEmbeddings, FlairEmbeddings, DocumentRNNEmbeddings
from flair.models import TextClassifier
from flair.trainers import ModelTrainer
from pathlib import Path
corpus = NLPTaskDataFetcher.load_classification_corpus(Path('./'), test_file='testform.txt', dev_file='devform.txt', train_file='trainform.txt')
word_embeddings = [WordEmbeddings('glove'), FlairEmbeddings('id-forward'), FlairEmbeddings('id-backward')]
document_embeddings = DocumentRNNEmbeddings(word_embeddings, hidden_size=256, bidirectional=True,rnn_type='lstm',rnn_layers=1,dropout=0.2)
classifier = TextClassifier(document_embeddings, label_dictionary=corpus.make_label_dictionary(), multi_label=False)
trainer = ModelTrainer(classifier, corpus)
trainer.train('./', max_epochs=30,checkpoint=True)
!mv best-model.pt best-form-model.pt
2019-07-17 14:01:26,730 Reading data from .
2019-07-17 14:01:26,733 Train: trainform.txt
2019-07-17 14:01:26,734 Dev: devform.txt
2019-07-17 14:01:26,735 Test: testform.txt
/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:7: DeprecationWarning: Call to deprecated function (or staticmethod) load_classification_corpus. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
import sys
/usr/local/lib/python3.6/dist-packages/flair/data_fetcher.py:447: DeprecationWarning: Call to deprecated function (or staticmethod) read_text_classification_file. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
max_tokens_per_doc=max_tokens_per_doc,
/usr/local/lib/python3.6/dist-packages/flair/data_fetcher.py:454: DeprecationWarning: Call to deprecated function (or staticmethod) read_text_classification_file. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
max_tokens_per_doc=max_tokens_per_doc,
/usr/local/lib/python3.6/dist-packages/flair/data_fetcher.py:463: DeprecationWarning: Call to deprecated function (or staticmethod) read_text_classification_file. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
max_tokens_per_doc=max_tokens_per_doc,
/usr/local/lib/python3.6/dist-packages/smart_open/smart_open_lib.py:398: UserWarning: This function is deprecated, use smart_open.open instead. See the migration notes for details: https://github.com/RaRe-Technologies/smart_open/blob/master/README.rst#migrating-to-the-new-open-function
'See the migration notes for details: %s' % _MIGRATION_NOTES_URL
2019-07-17 14:01:29,648 Computing label dictionary. Progress:
100%|██████████| 80/80 [00:00<00:00, 35409.91it/s]
2019-07-17 14:01:29,657 [b'0', b'1', b'2']
2019-07-17 14:01:29,876 ----------------------------------------------------------------------------------------------------
2019-07-17 14:01:29,877 Model: "TextClassifier(
(document_embeddings): DocumentRNNEmbeddings(
(embeddings): StackedEmbeddings(
(list_embedding_0): WordEmbeddings('glove')
(list_embedding_1): FlairEmbeddings(
(lm): LanguageModel(
(drop): Dropout(p=0.1)
(encoder): Embedding(8823, 100)
(rnn): LSTM(100, 2048)
(decoder): Linear(in_features=2048, out_features=8823, bias=True)
)
)
(list_embedding_2): FlairEmbeddings(
(lm): LanguageModel(
(drop): Dropout(p=0.1)
(encoder): Embedding(8823, 100)
(rnn): LSTM(100, 2048)
(decoder): Linear(in_features=2048, out_features=8823, bias=True)
)
)
)
(word_reprojection_map): Linear(in_features=4196, out_features=4196, bias=True)
(rnn): GRU(4196, 256, bidirectional=True)
(dropout): Dropout(p=0.2)
)
(decoder): Linear(in_features=1024, out_features=3, bias=True)
(loss_function): CrossEntropyLoss()
)"
2019-07-17 14:01:29,878 ----------------------------------------------------------------------------------------------------
2019-07-17 14:01:29,880 Corpus: "Corpus: 80 train + 20 dev + 257 test sentences"
2019-07-17 14:01:29,881 ----------------------------------------------------------------------------------------------------
2019-07-17 14:01:29,882 Parameters:
2019-07-17 14:01:29,884 - learning_rate: "0.1"
2019-07-17 14:01:29,887 - mini_batch_size: "32"
2019-07-17 14:01:29,888 - patience: "3"
2019-07-17 14:01:29,889 - anneal_factor: "0.5"
2019-07-17 14:01:29,891 - max_epochs: "30"
2019-07-17 14:01:29,892 - shuffle: "True"
2019-07-17 14:01:29,894 - train_with_dev: "False"
2019-07-17 14:01:29,896 ----------------------------------------------------------------------------------------------------
2019-07-17 14:01:29,898 Model training base path: "."
2019-07-17 14:01:29,899 ----------------------------------------------------------------------------------------------------
2019-07-17 14:01:29,901 Device: cuda:0
2019-07-17 14:01:29,903 ----------------------------------------------------------------------------------------------------
2019-07-17 14:01:29,905 Embedding storage mode: cpu
2019-07-17 14:01:29,907 ----------------------------------------------------------------------------------------------------
2019-07-17 14:01:30,520 epoch 1 - iter 0/3 - loss 1.09418356
2019-07-17 14:01:30,995 epoch 1 - iter 1/3 - loss 1.19802183
2019-07-17 14:01:31,416 epoch 1 - iter 2/3 - loss 1.03827896
2019-07-17 14:01:31,435 ----------------------------------------------------------------------------------------------------
2019-07-17 14:01:31,435 EPOCH 1 done: loss 1.0383 - lr 0.1000
2019-07-17 14:01:31,783 DEV : loss 1.3526480197906494 - score 0.55
2019-07-17 14:01:31,812 BAD EPOCHS (no improvement): 0
2019-07-17 14:01:44,230 ----------------------------------------------------------------------------------------------------
2019-07-17 14:01:44,338 epoch 2 - iter 0/3 - loss 0.34222460
2019-07-17 14:01:44,450 epoch 2 - iter 1/3 - loss 0.53306767
2019-07-17 14:01:44,509 epoch 2 - iter 2/3 - loss 0.50951793
2019-07-17 14:01:44,527 ----------------------------------------------------------------------------------------------------
2019-07-17 14:01:44,528 EPOCH 2 done: loss 0.5095 - lr 0.1000
2019-07-17 14:01:44,559 DEV : loss 1.1383740901947021 - score 0.55
2019-07-17 14:01:44,563 BAD EPOCHS (no improvement): 1
2019-07-17 14:01:56,141 ----------------------------------------------------------------------------------------------------
2019-07-17 14:01:56,244 epoch 3 - iter 0/3 - loss 0.35160568
2019-07-17 14:01:56,364 epoch 3 - iter 1/3 - loss 0.40961516
2019-07-17 14:01:56,457 epoch 3 - iter 2/3 - loss 0.55402919
2019-07-17 14:01:56,475 ----------------------------------------------------------------------------------------------------
2019-07-17 14:01:56,476 EPOCH 3 done: loss 0.5540 - lr 0.1000
2019-07-17 14:01:56,506 DEV : loss 1.1990362405776978 - score 0.6
2019-07-17 14:01:56,510 BAD EPOCHS (no improvement): 0
2019-07-17 14:02:08,134 ----------------------------------------------------------------------------------------------------
2019-07-17 14:02:08,240 epoch 4 - iter 0/3 - loss 0.30157968
2019-07-17 14:02:08,351 epoch 4 - iter 1/3 - loss 0.28370249
2019-07-17 14:02:08,419 epoch 4 - iter 2/3 - loss 0.32917794
2019-07-17 14:02:08,438 ----------------------------------------------------------------------------------------------------
2019-07-17 14:02:08,439 EPOCH 4 done: loss 0.3292 - lr 0.1000
2019-07-17 14:02:08,470 DEV : loss 2.735830545425415 - score 0.55
2019-07-17 14:02:08,473 BAD EPOCHS (no improvement): 1
2019-07-17 14:02:14,217 ----------------------------------------------------------------------------------------------------
2019-07-17 14:02:14,317 epoch 5 - iter 0/3 - loss 0.32484746
2019-07-17 14:02:14,449 epoch 5 - iter 1/3 - loss 0.37949546
2019-07-17 14:02:14,522 epoch 5 - iter 2/3 - loss 0.35487974
2019-07-17 14:02:14,542 ----------------------------------------------------------------------------------------------------
2019-07-17 14:02:14,543 EPOCH 5 done: loss 0.3549 - lr 0.1000
2019-07-17 14:02:14,574 DEV : loss 1.6706857681274414 - score 0.55
2019-07-17 14:02:14,577 BAD EPOCHS (no improvement): 2
2019-07-17 14:02:20,360 ----------------------------------------------------------------------------------------------------
2019-07-17 14:02:20,641 epoch 6 - iter 0/3 - loss 0.16572753
2019-07-17 14:02:20,758 epoch 6 - iter 1/3 - loss 0.16796360
2019-07-17 14:02:20,814 epoch 6 - iter 2/3 - loss 0.15461832
2019-07-17 14:02:20,831 ----------------------------------------------------------------------------------------------------
2019-07-17 14:02:20,832 EPOCH 6 done: loss 0.1546 - lr 0.1000
2019-07-17 14:02:20,860 DEV : loss 1.2553964853286743 - score 0.6
2019-07-17 14:02:20,864 BAD EPOCHS (no improvement): 3
2019-07-17 14:02:32,396 ----------------------------------------------------------------------------------------------------
2019-07-17 14:02:32,524 epoch 7 - iter 0/3 - loss 0.08774893
2019-07-17 14:02:32,631 epoch 7 - iter 1/3 - loss 0.13572774
2019-07-17 14:02:32,700 epoch 7 - iter 2/3 - loss 0.11693933
2019-07-17 14:02:32,721 ----------------------------------------------------------------------------------------------------
2019-07-17 14:02:32,722 EPOCH 7 done: loss 0.1169 - lr 0.1000
2019-07-17 14:02:32,755 DEV : loss 1.5748202800750732 - score 0.55
Epoch 6: reducing learning rate of group 0 to 5.0000e-02.
2019-07-17 14:02:32,759 BAD EPOCHS (no improvement): 4
2019-07-17 14:02:38,530 ----------------------------------------------------------------------------------------------------
2019-07-17 14:02:38,632 epoch 8 - iter 0/3 - loss 0.12741853
2019-07-17 14:02:38,759 epoch 8 - iter 1/3 - loss 0.09624327
2019-07-17 14:02:38,826 epoch 8 - iter 2/3 - loss 0.10591155
2019-07-17 14:02:38,844 ----------------------------------------------------------------------------------------------------
2019-07-17 14:02:38,845 EPOCH 8 done: loss 0.1059 - lr 0.0500
2019-07-17 14:02:38,876 DEV : loss 1.3528852462768555 - score 0.55
2019-07-17 14:02:38,881 BAD EPOCHS (no improvement): 1
2019-07-17 14:02:44,594 ----------------------------------------------------------------------------------------------------
2019-07-17 14:02:44,703 epoch 9 - iter 0/3 - loss 0.10527994
2019-07-17 14:02:44,825 epoch 9 - iter 1/3 - loss 0.09328366
2019-07-17 14:02:44,889 epoch 9 - iter 2/3 - loss 0.09436670
2019-07-17 14:02:44,908 ----------------------------------------------------------------------------------------------------
2019-07-17 14:02:44,909 EPOCH 9 done: loss 0.0944 - lr 0.0500
2019-07-17 14:02:44,940 DEV : loss 1.3984034061431885 - score 0.55
2019-07-17 14:02:44,944 BAD EPOCHS (no improvement): 2
2019-07-17 14:02:50,631 ----------------------------------------------------------------------------------------------------
2019-07-17 14:02:50,727 epoch 10 - iter 0/3 - loss 0.06243092
2019-07-17 14:02:50,868 epoch 10 - iter 1/3 - loss 0.07805193
2019-07-17 14:02:50,937 epoch 10 - iter 2/3 - loss 0.07599066
2019-07-17 14:02:50,955 ----------------------------------------------------------------------------------------------------
2019-07-17 14:02:50,956 EPOCH 10 done: loss 0.0760 - lr 0.0500
2019-07-17 14:02:50,989 DEV : loss 1.3679312467575073 - score 0.55
2019-07-17 14:02:50,994 BAD EPOCHS (no improvement): 3
2019-07-17 14:02:56,734 ----------------------------------------------------------------------------------------------------
2019-07-17 14:02:56,832 epoch 11 - iter 0/3 - loss 0.07850219
2019-07-17 14:02:56,955 epoch 11 - iter 1/3 - loss 0.06900330
2019-07-17 14:02:57,018 epoch 11 - iter 2/3 - loss 0.07356806
2019-07-17 14:02:57,036 ----------------------------------------------------------------------------------------------------
2019-07-17 14:02:57,037 EPOCH 11 done: loss 0.0736 - lr 0.0500
2019-07-17 14:02:57,069 DEV : loss 1.40873122215271 - score 0.55
Epoch 10: reducing learning rate of group 0 to 2.5000e-02.
2019-07-17 14:02:57,073 BAD EPOCHS (no improvement): 4
2019-07-17 14:03:02,993 ----------------------------------------------------------------------------------------------------
2019-07-17 14:03:03,094 epoch 12 - iter 0/3 - loss 0.06410529
2019-07-17 14:03:03,216 epoch 12 - iter 1/3 - loss 0.06309524
2019-07-17 14:03:03,276 epoch 12 - iter 2/3 - loss 0.06320149
2019-07-17 14:03:03,294 ----------------------------------------------------------------------------------------------------
2019-07-17 14:03:03,295 EPOCH 12 done: loss 0.0632 - lr 0.0250
2019-07-17 14:03:03,332 DEV : loss 1.4069684743881226 - score 0.6
2019-07-17 14:03:03,335 BAD EPOCHS (no improvement): 1
2019-07-17 14:03:14,835 ----------------------------------------------------------------------------------------------------
2019-07-17 14:03:14,942 epoch 13 - iter 0/3 - loss 0.08832344
2019-07-17 14:03:15,055 epoch 13 - iter 1/3 - loss 0.06960674
2019-07-17 14:03:15,120 epoch 13 - iter 2/3 - loss 0.06756547
2019-07-17 14:03:15,137 ----------------------------------------------------------------------------------------------------
2019-07-17 14:03:15,138 EPOCH 13 done: loss 0.0676 - lr 0.0250
2019-07-17 14:03:15,171 DEV : loss 1.4361594915390015 - score 0.55
2019-07-17 14:03:15,174 BAD EPOCHS (no improvement): 2
2019-07-17 14:03:21,039 ----------------------------------------------------------------------------------------------------
2019-07-17 14:03:21,144 epoch 14 - iter 0/3 - loss 0.07779028
2019-07-17 14:03:21,267 epoch 14 - iter 1/3 - loss 0.06480380
2019-07-17 14:03:21,333 epoch 14 - iter 2/3 - loss 0.06329538
2019-07-17 14:03:21,350 ----------------------------------------------------------------------------------------------------
2019-07-17 14:03:21,351 EPOCH 14 done: loss 0.0633 - lr 0.0250
2019-07-17 14:03:21,382 DEV : loss 1.4236245155334473 - score 0.6
2019-07-17 14:03:21,386 BAD EPOCHS (no improvement): 3
2019-07-17 14:03:33,001 ----------------------------------------------------------------------------------------------------
2019-07-17 14:03:33,171 epoch 15 - iter 0/3 - loss 0.05381881
2019-07-17 14:03:33,272 epoch 15 - iter 1/3 - loss 0.05450349
2019-07-17 14:03:33,331 epoch 15 - iter 2/3 - loss 0.06206007
2019-07-17 14:03:33,349 ----------------------------------------------------------------------------------------------------
2019-07-17 14:03:33,350 EPOCH 15 done: loss 0.0621 - lr 0.0250
2019-07-17 14:03:33,380 DEV : loss 1.4234994649887085 - score 0.6
Epoch 14: reducing learning rate of group 0 to 1.2500e-02.
2019-07-17 14:03:33,384 BAD EPOCHS (no improvement): 4
2019-07-17 14:03:45,015 ----------------------------------------------------------------------------------------------------
2019-07-17 14:03:45,256 epoch 16 - iter 0/3 - loss 0.05220055
2019-07-17 14:03:45,360 epoch 16 - iter 1/3 - loss 0.05680398
2019-07-17 14:03:45,420 epoch 16 - iter 2/3 - loss 0.05912224
2019-07-17 14:03:45,438 ----------------------------------------------------------------------------------------------------
2019-07-17 14:03:45,439 EPOCH 16 done: loss 0.0591 - lr 0.0125
2019-07-17 14:03:45,469 DEV : loss 1.4342578649520874 - score 0.6
2019-07-17 14:03:45,472 BAD EPOCHS (no improvement): 1
2019-07-17 14:03:57,263 ----------------------------------------------------------------------------------------------------
2019-07-17 14:03:57,446 epoch 17 - iter 0/3 - loss 0.05155280
2019-07-17 14:03:57,552 epoch 17 - iter 1/3 - loss 0.05187963
2019-07-17 14:03:57,610 epoch 17 - iter 2/3 - loss 0.05708090
2019-07-17 14:03:57,627 ----------------------------------------------------------------------------------------------------
2019-07-17 14:03:57,628 EPOCH 17 done: loss 0.0571 - lr 0.0125
2019-07-17 14:03:57,657 DEV : loss 1.4255011081695557 - score 0.6
2019-07-17 14:03:57,661 BAD EPOCHS (no improvement): 2
2019-07-17 14:04:09,239 ----------------------------------------------------------------------------------------------------
2019-07-17 14:04:09,421 epoch 18 - iter 0/3 - loss 0.07698785
2019-07-17 14:04:09,532 epoch 18 - iter 1/3 - loss 0.06120672
2019-07-17 14:04:09,608 epoch 18 - iter 2/3 - loss 0.05390206
2019-07-17 14:04:09,627 ----------------------------------------------------------------------------------------------------
2019-07-17 14:04:09,628 EPOCH 18 done: loss 0.0539 - lr 0.0125
2019-07-17 14:04:09,662 DEV : loss 1.4354857206344604 - score 0.6
2019-07-17 14:04:09,666 BAD EPOCHS (no improvement): 3
2019-07-17 14:04:21,464 ----------------------------------------------------------------------------------------------------
2019-07-17 14:04:21,691 epoch 19 - iter 0/3 - loss 0.05490704
2019-07-17 14:04:21,809 epoch 19 - iter 1/3 - loss 0.06230848
2019-07-17 14:04:21,876 epoch 19 - iter 2/3 - loss 0.05400377
2019-07-17 14:04:21,894 ----------------------------------------------------------------------------------------------------
2019-07-17 14:04:21,894 EPOCH 19 done: loss 0.0540 - lr 0.0125
2019-07-17 14:04:21,925 DEV : loss 1.4457045793533325 - score 0.6
Epoch 18: reducing learning rate of group 0 to 6.2500e-03.
2019-07-17 14:04:21,928 BAD EPOCHS (no improvement): 4
2019-07-17 14:04:33,421 ----------------------------------------------------------------------------------------------------
2019-07-17 14:04:33,709 epoch 20 - iter 0/3 - loss 0.04658061
2019-07-17 14:04:33,820 epoch 20 - iter 1/3 - loss 0.04831868
2019-07-17 14:04:33,889 epoch 20 - iter 2/3 - loss 0.05894887
2019-07-17 14:04:33,906 ----------------------------------------------------------------------------------------------------
2019-07-17 14:04:33,907 EPOCH 20 done: loss 0.0589 - lr 0.0063
2019-07-17 14:04:33,943 DEV : loss 1.4466190338134766 - score 0.6
2019-07-17 14:04:33,946 BAD EPOCHS (no improvement): 1
2019-07-17 14:04:45,785 ----------------------------------------------------------------------------------------------------
2019-07-17 14:04:45,975 epoch 21 - iter 0/3 - loss 0.04914344
2019-07-17 14:04:46,090 epoch 21 - iter 1/3 - loss 0.05831811
2019-07-17 14:04:46,165 epoch 21 - iter 2/3 - loss 0.05132634
2019-07-17 14:04:46,184 ----------------------------------------------------------------------------------------------------
2019-07-17 14:04:46,184 EPOCH 21 done: loss 0.0513 - lr 0.0063
2019-07-17 14:04:46,216 DEV : loss 1.4496076107025146 - score 0.6
2019-07-17 14:04:46,220 BAD EPOCHS (no improvement): 2
2019-07-17 14:04:58,505 ----------------------------------------------------------------------------------------------------
2019-07-17 14:04:58,618 epoch 22 - iter 0/3 - loss 0.05305767
2019-07-17 14:04:58,726 epoch 22 - iter 1/3 - loss 0.05053032
2019-07-17 14:04:58,800 epoch 22 - iter 2/3 - loss 0.05018285
2019-07-17 14:04:58,821 ----------------------------------------------------------------------------------------------------
2019-07-17 14:04:58,822 EPOCH 22 done: loss 0.0502 - lr 0.0063
2019-07-17 14:04:58,859 DEV : loss 1.4536397457122803 - score 0.6
2019-07-17 14:04:58,865 BAD EPOCHS (no improvement): 3
2019-07-17 14:05:10,396 ----------------------------------------------------------------------------------------------------
2019-07-17 14:05:10,728 epoch 23 - iter 0/3 - loss 0.04592466
2019-07-17 14:05:10,833 epoch 23 - iter 1/3 - loss 0.04925469
2019-07-17 14:05:10,900 epoch 23 - iter 2/3 - loss 0.04870697
2019-07-17 14:05:10,918 ----------------------------------------------------------------------------------------------------
2019-07-17 14:05:10,920 EPOCH 23 done: loss 0.0487 - lr 0.0063
2019-07-17 14:05:10,952 DEV : loss 1.45394766330719 - score 0.6
Epoch 22: reducing learning rate of group 0 to 3.1250e-03.
2019-07-17 14:05:10,958 BAD EPOCHS (no improvement): 4
2019-07-17 14:05:22,794 ----------------------------------------------------------------------------------------------------
2019-07-17 14:05:23,076 epoch 24 - iter 0/3 - loss 0.06067127
2019-07-17 14:05:23,187 epoch 24 - iter 1/3 - loss 0.05735061
2019-07-17 14:05:23,244 epoch 24 - iter 2/3 - loss 0.05255023
2019-07-17 14:05:23,263 ----------------------------------------------------------------------------------------------------
2019-07-17 14:05:23,265 EPOCH 24 done: loss 0.0526 - lr 0.0031
2019-07-17 14:05:23,297 DEV : loss 1.454268217086792 - score 0.6
2019-07-17 14:05:23,301 BAD EPOCHS (no improvement): 1
2019-07-17 14:05:34,955 ----------------------------------------------------------------------------------------------------
2019-07-17 14:05:35,266 epoch 25 - iter 0/3 - loss 0.05180541
2019-07-17 14:05:35,379 epoch 25 - iter 1/3 - loss 0.05030221
2019-07-17 14:05:35,443 epoch 25 - iter 2/3 - loss 0.05825821
2019-07-17 14:05:35,461 ----------------------------------------------------------------------------------------------------
2019-07-17 14:05:35,462 EPOCH 25 done: loss 0.0583 - lr 0.0031
2019-07-17 14:05:35,497 DEV : loss 1.4548426866531372 - score 0.6
2019-07-17 14:05:35,501 BAD EPOCHS (no improvement): 2
2019-07-17 14:05:47,020 ----------------------------------------------------------------------------------------------------
2019-07-17 14:05:47,335 epoch 26 - iter 0/3 - loss 0.05860512
2019-07-17 14:05:47,445 epoch 26 - iter 1/3 - loss 0.05079667
2019-07-17 14:05:47,511 epoch 26 - iter 2/3 - loss 0.05100415
2019-07-17 14:05:47,529 ----------------------------------------------------------------------------------------------------
2019-07-17 14:05:47,530 EPOCH 26 done: loss 0.0510 - lr 0.0031
2019-07-17 14:05:47,560 DEV : loss 1.4563053846359253 - score 0.6
2019-07-17 14:05:47,565 BAD EPOCHS (no improvement): 3
2019-07-17 14:05:59,317 ----------------------------------------------------------------------------------------------------
2019-07-17 14:05:59,521 epoch 27 - iter 0/3 - loss 0.03423696
2019-07-17 14:05:59,637 epoch 27 - iter 1/3 - loss 0.05334332
2019-07-17 14:05:59,706 epoch 27 - iter 2/3 - loss 0.05539700
2019-07-17 14:05:59,725 ----------------------------------------------------------------------------------------------------
2019-07-17 14:05:59,726 EPOCH 27 done: loss 0.0554 - lr 0.0031
2019-07-17 14:05:59,759 DEV : loss 1.4527745246887207 - score 0.6
Epoch 26: reducing learning rate of group 0 to 1.5625e-03.
2019-07-17 14:05:59,763 BAD EPOCHS (no improvement): 4
2019-07-17 14:06:11,307 ----------------------------------------------------------------------------------------------------
2019-07-17 14:06:11,685 epoch 28 - iter 0/3 - loss 0.04665977
2019-07-17 14:06:11,793 epoch 28 - iter 1/3 - loss 0.05624663
2019-07-17 14:06:11,851 epoch 28 - iter 2/3 - loss 0.05305390
2019-07-17 14:06:11,868 ----------------------------------------------------------------------------------------------------
2019-07-17 14:06:11,869 EPOCH 28 done: loss 0.0531 - lr 0.0016
2019-07-17 14:06:11,899 DEV : loss 1.4530184268951416 - score 0.6
2019-07-17 14:06:11,902 BAD EPOCHS (no improvement): 1
2019-07-17 14:06:23,341 ----------------------------------------------------------------------------------------------------
2019-07-17 14:06:23,635 epoch 29 - iter 0/3 - loss 0.04614327
2019-07-17 14:06:23,757 epoch 29 - iter 1/3 - loss 0.05186255
2019-07-17 14:06:23,824 epoch 29 - iter 2/3 - loss 0.05022394
2019-07-17 14:06:23,843 ----------------------------------------------------------------------------------------------------
2019-07-17 14:06:23,844 EPOCH 29 done: loss 0.0502 - lr 0.0016
2019-07-17 14:06:23,875 DEV : loss 1.453881025314331 - score 0.6
2019-07-17 14:06:23,878 BAD EPOCHS (no improvement): 2
2019-07-17 14:06:35,944 ----------------------------------------------------------------------------------------------------
2019-07-17 14:06:36,147 epoch 30 - iter 0/3 - loss 0.05481225
2019-07-17 14:06:36,265 epoch 30 - iter 1/3 - loss 0.05209364
2019-07-17 14:06:36,321 epoch 30 - iter 2/3 - loss 0.04601886
2019-07-17 14:06:36,340 ----------------------------------------------------------------------------------------------------
2019-07-17 14:06:36,341 EPOCH 30 done: loss 0.0460 - lr 0.0016
2019-07-17 14:06:36,372 DEV : loss 1.4555436372756958 - score 0.6
2019-07-17 14:06:36,375 BAD EPOCHS (no improvement): 3
2019-07-17 14:06:53,899 ----------------------------------------------------------------------------------------------------
2019-07-17 14:06:54,021 Testing using best model ...
2019-07-17 14:06:54,028 loading file best-model.pt
2019-07-17 14:07:00,731 0.7004 0.7004 0.7004
2019-07-17 14:07:00,732
MICRO_AVG: acc 0.5389 - f1-score 0.7004
MACRO_AVG: acc 0.2876 - f1-score 0.37116666666666664
0 tp: 167 - fp: 65 - fn: 11 - tn: 14 - precision: 0.7198 - recall: 0.9382 - accuracy: 0.6872 - f1-score: 0.8146
1 tp: 13 - fp: 12 - fn: 49 - tn: 183 - precision: 0.5200 - recall: 0.2097 - accuracy: 0.1757 - f1-score: 0.2989
2 tp: 0 - fp: 0 - fn: 17 - tn: 240 - precision: 0.0000 - recall: 0.0000 - accuracy: 0.0000 - f1-score: 0.0000
2019-07-17 14:07:00,734 ----------------------------------------------------------------------------------------------------
In [54]:
!ls
adc.json devkomm.txt testform.txt training.log
best-cred-model.pt devkoor.txt testkomm.txt trainkomm.txt
best-form-model.pt dev.txt testkoor.txt trainkoor.txt
checkpoint.pt final-model.pt test_sample.csv train.txt
corpus_rev_cleaned_sample.xlsx loss.tsv test.tsv weights.txt
corpus_rev_cleaned.xlsx sample_data test.txt
deveff.txt stopwordsid.txt traineff.txt
devform.txt testeff.txt trainform.txt
In [0]:
import os
import math
import random
# os.remove("traineff.txt")
# os.remove("testeff.txt")
# os.remove("deveff.txt")
listrand = list(range(100))
random.shuffle(listrand)
listranddev = []
for i in range(math.floor(len(listrand)*0.2)):
listranddev.append(listrand.pop())
file1 = open("traineff.txt","a", encoding="utf-8")
for i in listrand:
file1.writelines(['__label__',str(dfcred['efektivitas'].iloc[i]),' ',str(dfcred['sentence'].iloc[i]),'\n'])
file1.close()
file2 = open("deveff.txt","a", encoding="utf-8")
for i in listranddev:
file2.writelines(['__label__',str(dfcred['efektivitas'].iloc[i]),' ',str(dfcred['sentence'].iloc[i]),'\n'])
file2.close()
file3 = open("testeff.txt","a", encoding="utf-8")
for i in range(len(dftestcred)):
file3.writelines(['__label__',str(dftestcred['efektivitas'].iloc[i]),' ',str(dftestcred['sentence'].iloc[i]),'\n'])
file3.close()
In [55]:
from flair.data_fetcher import NLPTaskDataFetcher
from flair.embeddings import WordEmbeddings, FlairEmbeddings, DocumentRNNEmbeddings
from flair.models import TextClassifier
from flair.trainers import ModelTrainer
from pathlib import Path
corpus = NLPTaskDataFetcher.load_classification_corpus(Path('./'), test_file='testeff.txt', dev_file='deveff.txt', train_file='traineff.txt')
word_embeddings = [WordEmbeddings('glove'), FlairEmbeddings('id-forward'), FlairEmbeddings('id-backward')]
document_embeddings = DocumentRNNEmbeddings(word_embeddings, hidden_size=256, bidirectional=True,rnn_type='lstm',rnn_layers=1,dropout=0.2)
classifier = TextClassifier(document_embeddings, label_dictionary=corpus.make_label_dictionary(), multi_label=False)
trainer = ModelTrainer(classifier, corpus)
trainer.train('./', max_epochs=30,checkpoint=True)
!mv best-model.pt best-eff-model.pt
2019-07-17 14:08:01,992 Reading data from .
2019-07-17 14:08:01,993 Train: traineff.txt
2019-07-17 14:08:01,995 Dev: deveff.txt
2019-07-17 14:08:01,998 Test: testeff.txt
/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:7: DeprecationWarning: Call to deprecated function (or staticmethod) load_classification_corpus. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
import sys
/usr/local/lib/python3.6/dist-packages/flair/data_fetcher.py:447: DeprecationWarning: Call to deprecated function (or staticmethod) read_text_classification_file. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
max_tokens_per_doc=max_tokens_per_doc,
/usr/local/lib/python3.6/dist-packages/flair/data_fetcher.py:454: DeprecationWarning: Call to deprecated function (or staticmethod) read_text_classification_file. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
max_tokens_per_doc=max_tokens_per_doc,
/usr/local/lib/python3.6/dist-packages/flair/data_fetcher.py:463: DeprecationWarning: Call to deprecated function (or staticmethod) read_text_classification_file. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
max_tokens_per_doc=max_tokens_per_doc,
/usr/local/lib/python3.6/dist-packages/smart_open/smart_open_lib.py:398: UserWarning: This function is deprecated, use smart_open.open instead. See the migration notes for details: https://github.com/RaRe-Technologies/smart_open/blob/master/README.rst#migrating-to-the-new-open-function
'See the migration notes for details: %s' % _MIGRATION_NOTES_URL
2019-07-17 14:08:05,267 Computing label dictionary. Progress:
100%|██████████| 80/80 [00:00<00:00, 28469.74it/s]
2019-07-17 14:08:05,276 [b'1', b'0', b'2']
2019-07-17 14:08:05,513 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:05,516 Model: "TextClassifier(
(document_embeddings): DocumentRNNEmbeddings(
(embeddings): StackedEmbeddings(
(list_embedding_0): WordEmbeddings('glove')
(list_embedding_1): FlairEmbeddings(
(lm): LanguageModel(
(drop): Dropout(p=0.1)
(encoder): Embedding(8823, 100)
(rnn): LSTM(100, 2048)
(decoder): Linear(in_features=2048, out_features=8823, bias=True)
)
)
(list_embedding_2): FlairEmbeddings(
(lm): LanguageModel(
(drop): Dropout(p=0.1)
(encoder): Embedding(8823, 100)
(rnn): LSTM(100, 2048)
(decoder): Linear(in_features=2048, out_features=8823, bias=True)
)
)
)
(word_reprojection_map): Linear(in_features=4196, out_features=4196, bias=True)
(rnn): GRU(4196, 256, bidirectional=True)
(dropout): Dropout(p=0.2)
)
(decoder): Linear(in_features=1024, out_features=3, bias=True)
(loss_function): CrossEntropyLoss()
)"
2019-07-17 14:08:05,518 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:05,519 Corpus: "Corpus: 80 train + 20 dev + 257 test sentences"
2019-07-17 14:08:05,521 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:05,523 Parameters:
2019-07-17 14:08:05,525 - learning_rate: "0.1"
2019-07-17 14:08:05,527 - mini_batch_size: "32"
2019-07-17 14:08:05,528 - patience: "3"
2019-07-17 14:08:05,529 - anneal_factor: "0.5"
2019-07-17 14:08:05,531 - max_epochs: "30"
2019-07-17 14:08:05,532 - shuffle: "True"
2019-07-17 14:08:05,534 - train_with_dev: "False"
2019-07-17 14:08:05,535 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:05,537 Model training base path: "."
2019-07-17 14:08:05,539 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:05,540 Device: cuda:0
2019-07-17 14:08:05,541 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:05,543 Embedding storage mode: cpu
2019-07-17 14:08:05,545 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:06,106 epoch 1 - iter 0/3 - loss 1.16303730
2019-07-17 14:08:06,646 epoch 1 - iter 1/3 - loss 1.64200783
2019-07-17 14:08:07,067 epoch 1 - iter 2/3 - loss 1.56402437
2019-07-17 14:08:07,085 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:07,086 EPOCH 1 done: loss 1.5640 - lr 0.1000
2019-07-17 14:08:07,368 DEV : loss 1.5192286968231201 - score 0.6
2019-07-17 14:08:07,396 BAD EPOCHS (no improvement): 0
2019-07-17 14:08:19,426 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:19,549 epoch 2 - iter 0/3 - loss 1.21964073
2019-07-17 14:08:19,665 epoch 2 - iter 1/3 - loss 1.05310905
2019-07-17 14:08:19,736 epoch 2 - iter 2/3 - loss 1.13709883
2019-07-17 14:08:19,757 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:19,757 EPOCH 2 done: loss 1.1371 - lr 0.1000
2019-07-17 14:08:19,789 DEV : loss 1.2374778985977173 - score 0.35
2019-07-17 14:08:19,794 BAD EPOCHS (no improvement): 1
2019-07-17 14:08:25,850 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:25,955 epoch 3 - iter 0/3 - loss 0.76153761
2019-07-17 14:08:26,080 epoch 3 - iter 1/3 - loss 0.76696792
2019-07-17 14:08:26,148 epoch 3 - iter 2/3 - loss 0.71887944
2019-07-17 14:08:26,166 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:26,167 EPOCH 3 done: loss 0.7189 - lr 0.1000
2019-07-17 14:08:26,199 DEV : loss 1.0521771907806396 - score 0.55
2019-07-17 14:08:26,203 BAD EPOCHS (no improvement): 2
2019-07-17 14:08:32,021 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:32,129 epoch 4 - iter 0/3 - loss 0.28146386
2019-07-17 14:08:32,251 epoch 4 - iter 1/3 - loss 0.37050270
2019-07-17 14:08:32,318 epoch 4 - iter 2/3 - loss 0.35931467
2019-07-17 14:08:32,337 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:32,338 EPOCH 4 done: loss 0.3593 - lr 0.1000
2019-07-17 14:08:32,368 DEV : loss 1.1522130966186523 - score 0.6
2019-07-17 14:08:32,371 BAD EPOCHS (no improvement): 3
2019-07-17 14:08:44,194 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:44,292 epoch 5 - iter 0/3 - loss 0.24514717
2019-07-17 14:08:44,409 epoch 5 - iter 1/3 - loss 0.26025668
2019-07-17 14:08:44,477 epoch 5 - iter 2/3 - loss 0.28091046
2019-07-17 14:08:44,495 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:44,496 EPOCH 5 done: loss 0.2809 - lr 0.1000
2019-07-17 14:08:44,523 DEV : loss 1.1429669857025146 - score 0.45
Epoch 4: reducing learning rate of group 0 to 5.0000e-02.
2019-07-17 14:08:44,526 BAD EPOCHS (no improvement): 4
2019-07-17 14:08:50,132 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:50,238 epoch 6 - iter 0/3 - loss 0.21652359
2019-07-17 14:08:50,371 epoch 6 - iter 1/3 - loss 0.21237388
2019-07-17 14:08:50,436 epoch 6 - iter 2/3 - loss 0.20616512
2019-07-17 14:08:50,454 ----------------------------------------------------------------------------------------------------
2019-07-17 14:08:50,455 EPOCH 6 done: loss 0.2062 - lr 0.0500
2019-07-17 14:08:50,484 DEV : loss 1.1216078996658325 - score 0.6
2019-07-17 14:08:50,487 BAD EPOCHS (no improvement): 1
2019-07-17 14:09:02,286 ----------------------------------------------------------------------------------------------------
2019-07-17 14:09:02,396 epoch 7 - iter 0/3 - loss 0.17846069
2019-07-17 14:09:02,525 epoch 7 - iter 1/3 - loss 0.16920908
2019-07-17 14:09:02,592 epoch 7 - iter 2/3 - loss 0.17669334
2019-07-17 14:09:02,610 ----------------------------------------------------------------------------------------------------
2019-07-17 14:09:02,611 EPOCH 7 done: loss 0.1767 - lr 0.0500
2019-07-17 14:09:02,640 DEV : loss 1.1476935148239136 - score 0.6
2019-07-17 14:09:02,643 BAD EPOCHS (no improvement): 2
2019-07-17 14:09:14,350 ----------------------------------------------------------------------------------------------------
2019-07-17 14:09:14,484 epoch 8 - iter 0/3 - loss 0.17236343
2019-07-17 14:09:14,595 epoch 8 - iter 1/3 - loss 0.15557738
2019-07-17 14:09:14,656 epoch 8 - iter 2/3 - loss 0.15871922
2019-07-17 14:09:14,679 ----------------------------------------------------------------------------------------------------
2019-07-17 14:09:14,680 EPOCH 8 done: loss 0.1587 - lr 0.0500
2019-07-17 14:09:14,710 DEV : loss 1.1637150049209595 - score 0.65
2019-07-17 14:09:14,716 BAD EPOCHS (no improvement): 0
2019-07-17 14:09:26,381 ----------------------------------------------------------------------------------------------------
2019-07-17 14:09:26,485 epoch 9 - iter 0/3 - loss 0.11427464
2019-07-17 14:09:26,599 epoch 9 - iter 1/3 - loss 0.12512656
2019-07-17 14:09:26,661 epoch 9 - iter 2/3 - loss 0.14061830
2019-07-17 14:09:26,679 ----------------------------------------------------------------------------------------------------
2019-07-17 14:09:26,680 EPOCH 9 done: loss 0.1406 - lr 0.0500
2019-07-17 14:09:26,708 DEV : loss 1.178027868270874 - score 0.6
2019-07-17 14:09:26,713 BAD EPOCHS (no improvement): 1
2019-07-17 14:09:32,432 ----------------------------------------------------------------------------------------------------
2019-07-17 14:09:32,533 epoch 10 - iter 0/3 - loss 0.12006348
2019-07-17 14:09:32,651 epoch 10 - iter 1/3 - loss 0.11882598
2019-07-17 14:09:32,725 epoch 10 - iter 2/3 - loss 0.12494637
2019-07-17 14:09:32,744 ----------------------------------------------------------------------------------------------------
2019-07-17 14:09:32,745 EPOCH 10 done: loss 0.1249 - lr 0.0500
2019-07-17 14:09:32,771 DEV : loss 1.19778311252594 - score 0.6
2019-07-17 14:09:32,774 BAD EPOCHS (no improvement): 2
2019-07-17 14:09:38,390 ----------------------------------------------------------------------------------------------------
2019-07-17 14:09:38,483 epoch 11 - iter 0/3 - loss 0.11437646
2019-07-17 14:09:38,601 epoch 11 - iter 1/3 - loss 0.10541795
2019-07-17 14:09:38,697 epoch 11 - iter 2/3 - loss 0.11076359
2019-07-17 14:09:38,715 ----------------------------------------------------------------------------------------------------
2019-07-17 14:09:38,715 EPOCH 11 done: loss 0.1108 - lr 0.0500
2019-07-17 14:09:38,744 DEV : loss 1.2188024520874023 - score 0.6
2019-07-17 14:09:38,747 BAD EPOCHS (no improvement): 3
2019-07-17 14:09:44,418 ----------------------------------------------------------------------------------------------------
2019-07-17 14:09:44,516 epoch 12 - iter 0/3 - loss 0.10127354
2019-07-17 14:09:44,658 epoch 12 - iter 1/3 - loss 0.09740097
2019-07-17 14:09:44,724 epoch 12 - iter 2/3 - loss 0.09797648
2019-07-17 14:09:44,742 ----------------------------------------------------------------------------------------------------
2019-07-17 14:09:44,743 EPOCH 12 done: loss 0.0980 - lr 0.0500
2019-07-17 14:09:44,770 DEV : loss 1.2249811887741089 - score 0.6
Epoch 11: reducing learning rate of group 0 to 2.5000e-02.
2019-07-17 14:09:44,773 BAD EPOCHS (no improvement): 4
2019-07-17 14:09:50,382 ----------------------------------------------------------------------------------------------------
2019-07-17 14:09:50,483 epoch 13 - iter 0/3 - loss 0.08229906
2019-07-17 14:09:50,603 epoch 13 - iter 1/3 - loss 0.08862678
2019-07-17 14:09:50,664 epoch 13 - iter 2/3 - loss 0.09050603
2019-07-17 14:09:50,682 ----------------------------------------------------------------------------------------------------
2019-07-17 14:09:50,683 EPOCH 13 done: loss 0.0905 - lr 0.0250
2019-07-17 14:09:50,715 DEV : loss 1.237436056137085 - score 0.6
2019-07-17 14:09:50,718 BAD EPOCHS (no improvement): 1
2019-07-17 14:09:56,419 ----------------------------------------------------------------------------------------------------
2019-07-17 14:09:56,523 epoch 14 - iter 0/3 - loss 0.09800007
2019-07-17 14:09:56,641 epoch 14 - iter 1/3 - loss 0.08687920
2019-07-17 14:09:56,741 epoch 14 - iter 2/3 - loss 0.08599245
2019-07-17 14:09:56,759 ----------------------------------------------------------------------------------------------------
2019-07-17 14:09:56,760 EPOCH 14 done: loss 0.0860 - lr 0.0250
2019-07-17 14:09:56,790 DEV : loss 1.2467365264892578 - score 0.65
2019-07-17 14:09:56,793 BAD EPOCHS (no improvement): 2
2019-07-17 14:10:08,091 ----------------------------------------------------------------------------------------------------
2019-07-17 14:10:08,190 epoch 15 - iter 0/3 - loss 0.08531683
2019-07-17 14:10:08,325 epoch 15 - iter 1/3 - loss 0.07991619
2019-07-17 14:10:08,394 epoch 15 - iter 2/3 - loss 0.08589431
2019-07-17 14:10:08,411 ----------------------------------------------------------------------------------------------------
2019-07-17 14:10:08,412 EPOCH 15 done: loss 0.0859 - lr 0.0250
2019-07-17 14:10:08,440 DEV : loss 1.2577483654022217 - score 0.65
2019-07-17 14:10:08,443 BAD EPOCHS (no improvement): 3
2019-07-17 14:10:20,166 ----------------------------------------------------------------------------------------------------
2019-07-17 14:10:20,351 epoch 16 - iter 0/3 - loss 0.07960413
2019-07-17 14:10:20,460 epoch 16 - iter 1/3 - loss 0.08036118
2019-07-17 14:10:20,517 epoch 16 - iter 2/3 - loss 0.07739619
2019-07-17 14:10:20,535 ----------------------------------------------------------------------------------------------------
2019-07-17 14:10:20,536 EPOCH 16 done: loss 0.0774 - lr 0.0250
2019-07-17 14:10:20,562 DEV : loss 1.2700064182281494 - score 0.6
Epoch 15: reducing learning rate of group 0 to 1.2500e-02.
2019-07-17 14:10:20,565 BAD EPOCHS (no improvement): 4
2019-07-17 14:10:26,340 ----------------------------------------------------------------------------------------------------
2019-07-17 14:10:26,443 epoch 17 - iter 0/3 - loss 0.07697551
2019-07-17 14:10:26,572 epoch 17 - iter 1/3 - loss 0.07759785
2019-07-17 14:10:26,652 epoch 17 - iter 2/3 - loss 0.07712884
2019-07-17 14:10:26,670 ----------------------------------------------------------------------------------------------------
2019-07-17 14:10:26,671 EPOCH 17 done: loss 0.0771 - lr 0.0125
2019-07-17 14:10:26,700 DEV : loss 1.2695549726486206 - score 0.6
2019-07-17 14:10:26,705 BAD EPOCHS (no improvement): 1
2019-07-17 14:10:32,633 ----------------------------------------------------------------------------------------------------
2019-07-17 14:10:32,736 epoch 18 - iter 0/3 - loss 0.07837548
2019-07-17 14:10:32,853 epoch 18 - iter 1/3 - loss 0.08127649
2019-07-17 14:10:32,922 epoch 18 - iter 2/3 - loss 0.07352381
2019-07-17 14:10:32,942 ----------------------------------------------------------------------------------------------------
2019-07-17 14:10:32,943 EPOCH 18 done: loss 0.0735 - lr 0.0125
2019-07-17 14:10:32,971 DEV : loss 1.274285912513733 - score 0.6
2019-07-17 14:10:32,974 BAD EPOCHS (no improvement): 2
2019-07-17 14:10:38,649 ----------------------------------------------------------------------------------------------------
2019-07-17 14:10:38,752 epoch 19 - iter 0/3 - loss 0.07476106
2019-07-17 14:10:38,870 epoch 19 - iter 1/3 - loss 0.07318940
2019-07-17 14:10:38,954 epoch 19 - iter 2/3 - loss 0.07598561
2019-07-17 14:10:38,971 ----------------------------------------------------------------------------------------------------
2019-07-17 14:10:38,972 EPOCH 19 done: loss 0.0760 - lr 0.0125
2019-07-17 14:10:39,002 DEV : loss 1.2799052000045776 - score 0.6
2019-07-17 14:10:39,005 BAD EPOCHS (no improvement): 3
2019-07-17 14:10:44,941 ----------------------------------------------------------------------------------------------------
2019-07-17 14:10:45,048 epoch 20 - iter 0/3 - loss 0.06224165
2019-07-17 14:10:45,155 epoch 20 - iter 1/3 - loss 0.06917006
2019-07-17 14:10:45,213 epoch 20 - iter 2/3 - loss 0.07199000
2019-07-17 14:10:45,231 ----------------------------------------------------------------------------------------------------
2019-07-17 14:10:45,232 EPOCH 20 done: loss 0.0720 - lr 0.0125
2019-07-17 14:10:45,261 DEV : loss 1.2800954580307007 - score 0.65
Epoch 19: reducing learning rate of group 0 to 6.2500e-03.
2019-07-17 14:10:45,264 BAD EPOCHS (no improvement): 4
2019-07-17 14:10:56,775 ----------------------------------------------------------------------------------------------------
2019-07-17 14:10:56,900 epoch 21 - iter 0/3 - loss 0.07532974
2019-07-17 14:10:57,027 epoch 21 - iter 1/3 - loss 0.07057948
2019-07-17 14:10:57,098 epoch 21 - iter 2/3 - loss 0.06772803
2019-07-17 14:10:57,119 ----------------------------------------------------------------------------------------------------
2019-07-17 14:10:57,120 EPOCH 21 done: loss 0.0677 - lr 0.0063
2019-07-17 14:10:57,151 DEV : loss 1.2829930782318115 - score 0.6
2019-07-17 14:10:57,155 BAD EPOCHS (no improvement): 1
2019-07-17 14:11:03,454 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:03,718 epoch 22 - iter 0/3 - loss 0.06762016
2019-07-17 14:11:03,824 epoch 22 - iter 1/3 - loss 0.07037774
2019-07-17 14:11:03,885 epoch 22 - iter 2/3 - loss 0.06812993
2019-07-17 14:11:03,902 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:03,903 EPOCH 22 done: loss 0.0681 - lr 0.0063
2019-07-17 14:11:03,929 DEV : loss 1.2856279611587524 - score 0.6
2019-07-17 14:11:03,932 BAD EPOCHS (no improvement): 2
2019-07-17 14:11:09,716 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:09,826 epoch 23 - iter 0/3 - loss 0.07128040
2019-07-17 14:11:09,956 epoch 23 - iter 1/3 - loss 0.06893324
2019-07-17 14:11:10,025 epoch 23 - iter 2/3 - loss 0.06872451
2019-07-17 14:11:10,043 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:10,043 EPOCH 23 done: loss 0.0687 - lr 0.0063
2019-07-17 14:11:10,075 DEV : loss 1.288386583328247 - score 0.6
2019-07-17 14:11:10,078 BAD EPOCHS (no improvement): 3
2019-07-17 14:11:15,715 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:15,820 epoch 24 - iter 0/3 - loss 0.08124591
2019-07-17 14:11:15,958 epoch 24 - iter 1/3 - loss 0.07520661
2019-07-17 14:11:16,023 epoch 24 - iter 2/3 - loss 0.06658916
2019-07-17 14:11:16,042 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:16,043 EPOCH 24 done: loss 0.0666 - lr 0.0063
2019-07-17 14:11:16,073 DEV : loss 1.2910805940628052 - score 0.6
Epoch 23: reducing learning rate of group 0 to 3.1250e-03.
2019-07-17 14:11:16,076 BAD EPOCHS (no improvement): 4
2019-07-17 14:11:21,832 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:21,936 epoch 25 - iter 0/3 - loss 0.06948145
2019-07-17 14:11:22,057 epoch 25 - iter 1/3 - loss 0.06901116
2019-07-17 14:11:22,122 epoch 25 - iter 2/3 - loss 0.06582189
2019-07-17 14:11:22,140 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:22,141 EPOCH 25 done: loss 0.0658 - lr 0.0031
2019-07-17 14:11:22,169 DEV : loss 1.2919700145721436 - score 0.6
2019-07-17 14:11:22,173 BAD EPOCHS (no improvement): 1
2019-07-17 14:11:27,998 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:28,095 epoch 26 - iter 0/3 - loss 0.07192482
2019-07-17 14:11:28,235 epoch 26 - iter 1/3 - loss 0.07184860
2019-07-17 14:11:28,308 epoch 26 - iter 2/3 - loss 0.06392838
2019-07-17 14:11:28,326 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:28,328 EPOCH 26 done: loss 0.0639 - lr 0.0031
2019-07-17 14:11:28,358 DEV : loss 1.2935829162597656 - score 0.6
2019-07-17 14:11:28,361 BAD EPOCHS (no improvement): 2
2019-07-17 14:11:34,068 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:34,180 epoch 27 - iter 0/3 - loss 0.06034578
2019-07-17 14:11:34,314 epoch 27 - iter 1/3 - loss 0.07480467
2019-07-17 14:11:34,385 epoch 27 - iter 2/3 - loss 0.07033366
2019-07-17 14:11:34,403 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:34,404 EPOCH 27 done: loss 0.0703 - lr 0.0031
2019-07-17 14:11:34,434 DEV : loss 1.293498158454895 - score 0.6
2019-07-17 14:11:34,437 BAD EPOCHS (no improvement): 3
2019-07-17 14:11:40,258 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:40,356 epoch 28 - iter 0/3 - loss 0.08878533
2019-07-17 14:11:40,480 epoch 28 - iter 1/3 - loss 0.07716479
2019-07-17 14:11:40,554 epoch 28 - iter 2/3 - loss 0.06764207
2019-07-17 14:11:40,574 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:40,575 EPOCH 28 done: loss 0.0676 - lr 0.0031
2019-07-17 14:11:40,606 DEV : loss 1.2948716878890991 - score 0.6
Epoch 27: reducing learning rate of group 0 to 1.5625e-03.
2019-07-17 14:11:40,611 BAD EPOCHS (no improvement): 4
2019-07-17 14:11:46,429 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:46,532 epoch 29 - iter 0/3 - loss 0.06877460
2019-07-17 14:11:46,655 epoch 29 - iter 1/3 - loss 0.06695824
2019-07-17 14:11:46,718 epoch 29 - iter 2/3 - loss 0.06814602
2019-07-17 14:11:46,736 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:46,737 EPOCH 29 done: loss 0.0681 - lr 0.0016
2019-07-17 14:11:46,766 DEV : loss 1.2957552671432495 - score 0.6
2019-07-17 14:11:46,769 BAD EPOCHS (no improvement): 1
2019-07-17 14:11:52,474 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:52,569 epoch 30 - iter 0/3 - loss 0.07509648
2019-07-17 14:11:52,700 epoch 30 - iter 1/3 - loss 0.06588296
2019-07-17 14:11:52,769 epoch 30 - iter 2/3 - loss 0.06841478
2019-07-17 14:11:52,788 ----------------------------------------------------------------------------------------------------
2019-07-17 14:11:52,789 EPOCH 30 done: loss 0.0684 - lr 0.0016
2019-07-17 14:11:52,819 DEV : loss 1.2959991693496704 - score 0.6
2019-07-17 14:11:52,823 BAD EPOCHS (no improvement): 2
2019-07-17 14:12:04,239 ----------------------------------------------------------------------------------------------------
2019-07-17 14:12:04,418 Testing using best model ...
2019-07-17 14:12:04,421 loading file best-model.pt
2019-07-17 14:12:14,538 0.677 0.677 0.677
2019-07-17 14:12:14,539
MICRO_AVG: acc 0.5118 - f1-score 0.677
MACRO_AVG: acc 0.243 - f1-score 0.304
0 tp: 170 - fp: 82 - fn: 1 - tn: 4 - precision: 0.6746 - recall: 0.9942 - accuracy: 0.6719 - f1-score: 0.8038
1 tp: 4 - fp: 1 - fn: 65 - tn: 187 - precision: 0.8000 - recall: 0.0580 - accuracy: 0.0571 - f1-score: 0.1082
2 tp: 0 - fp: 0 - fn: 17 - tn: 240 - precision: 0.0000 - recall: 0.0000 - accuracy: 0.0000 - f1-score: 0.0000
2019-07-17 14:12:14,542 ----------------------------------------------------------------------------------------------------
In [0]:
import os
import math
import random
# os.remove("trainkoor.txt")
# os.remove("testkoor.txt")
# os.remove("devkoor.txt")
listrand = list(range(100))
random.shuffle(listrand)
listranddev = []
for i in range(math.floor(len(listrand)*0.2)):
listranddev.append(listrand.pop())
file1 = open("trainkoor.txt","a", encoding="utf-8")
for i in listrand:
file1.writelines(['__label__',str(dfcred['koordinasi'].iloc[i]),' ',str(dfcred['sentence'].iloc[i]),'\n'])
file1.close()
file2 = open("devkoor.txt","a", encoding="utf-8")
for i in listranddev:
file2.writelines(['__label__',str(dfcred['koordinasi'].iloc[i]),' ',str(dfcred['sentence'].iloc[i]),'\n'])
file2.close()
file3 = open("testkoor.txt","a", encoding="utf-8")
for i in range(len(dftestcred)):
file3.writelines(['__label__',str(dftestcred['koordinasi'].iloc[i]),' ',str(dftestcred['sentence'].iloc[i]),'\n'])
file3.close()
In [46]:
from flair.data_fetcher import NLPTaskDataFetcher
from flair.embeddings import WordEmbeddings, FlairEmbeddings, DocumentRNNEmbeddings
from flair.models import TextClassifier
from flair.trainers import ModelTrainer
from pathlib import Path
corpus = NLPTaskDataFetcher.load_classification_corpus(Path('./'), test_file='testkoor.txt', dev_file='devkoor.txt', train_file='trainkoor.txt')
word_embeddings = [WordEmbeddings('glove'), FlairEmbeddings('id-forward'), FlairEmbeddings('id-backward')]
document_embeddings = DocumentRNNEmbeddings(word_embeddings, hidden_size=256, bidirectional=True,rnn_type='lstm',rnn_layers=1,dropout=0.2)
classifier = TextClassifier(document_embeddings, label_dictionary=corpus.make_label_dictionary(), multi_label=False)
trainer = ModelTrainer(classifier, corpus)
trainer.train('./', max_epochs=30,checkpoint=True)
!mv best-model.pt best-koor-model.pt
2019-07-17 13:42:13,492 Reading data from .
2019-07-17 13:42:13,494 Train: trainkoor.txt
2019-07-17 13:42:13,496 Dev: devkoor.txt
2019-07-17 13:42:13,497 Test: testkoor.txt
/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:7: DeprecationWarning: Call to deprecated function (or staticmethod) load_classification_corpus. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
import sys
/usr/local/lib/python3.6/dist-packages/flair/data_fetcher.py:447: DeprecationWarning: Call to deprecated function (or staticmethod) read_text_classification_file. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
max_tokens_per_doc=max_tokens_per_doc,
/usr/local/lib/python3.6/dist-packages/flair/data_fetcher.py:454: DeprecationWarning: Call to deprecated function (or staticmethod) read_text_classification_file. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
max_tokens_per_doc=max_tokens_per_doc,
/usr/local/lib/python3.6/dist-packages/flair/data_fetcher.py:463: DeprecationWarning: Call to deprecated function (or staticmethod) read_text_classification_file. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
max_tokens_per_doc=max_tokens_per_doc,
/usr/local/lib/python3.6/dist-packages/smart_open/smart_open_lib.py:398: UserWarning: This function is deprecated, use smart_open.open instead. See the migration notes for details: https://github.com/RaRe-Technologies/smart_open/blob/master/README.rst#migrating-to-the-new-open-function
'See the migration notes for details: %s' % _MIGRATION_NOTES_URL
2019-07-17 13:42:16,124 Computing label dictionary. Progress:
100%|██████████| 80/80 [00:00<00:00, 44804.96it/s]
2019-07-17 13:42:16,135 [b'0', b'1', b'2']
2019-07-17 13:42:16,290 ----------------------------------------------------------------------------------------------------
2019-07-17 13:42:16,291 Model: "TextClassifier(
(document_embeddings): DocumentRNNEmbeddings(
(embeddings): StackedEmbeddings(
(list_embedding_0): WordEmbeddings('glove')
(list_embedding_1): FlairEmbeddings(
(lm): LanguageModel(
(drop): Dropout(p=0.1)
(encoder): Embedding(8823, 100)
(rnn): LSTM(100, 2048)
(decoder): Linear(in_features=2048, out_features=8823, bias=True)
)
)
(list_embedding_2): FlairEmbeddings(
(lm): LanguageModel(
(drop): Dropout(p=0.1)
(encoder): Embedding(8823, 100)
(rnn): LSTM(100, 2048)
(decoder): Linear(in_features=2048, out_features=8823, bias=True)
)
)
)
(word_reprojection_map): Linear(in_features=4196, out_features=4196, bias=True)
(rnn): GRU(4196, 256, bidirectional=True)
(dropout): Dropout(p=0.2)
)
(decoder): Linear(in_features=1024, out_features=3, bias=True)
(loss_function): CrossEntropyLoss()
)"
2019-07-17 13:42:16,292 ----------------------------------------------------------------------------------------------------
2019-07-17 13:42:16,294 Corpus: "Corpus: 80 train + 20 dev + 257 test sentences"
2019-07-17 13:42:16,296 ----------------------------------------------------------------------------------------------------
2019-07-17 13:42:16,297 Parameters:
2019-07-17 13:42:16,299 - learning_rate: "0.1"
2019-07-17 13:42:16,300 - mini_batch_size: "32"
2019-07-17 13:42:16,301 - patience: "3"
2019-07-17 13:42:16,303 - anneal_factor: "0.5"
2019-07-17 13:42:16,305 - max_epochs: "30"
2019-07-17 13:42:16,307 - shuffle: "True"
2019-07-17 13:42:16,308 - train_with_dev: "False"
2019-07-17 13:42:16,310 ----------------------------------------------------------------------------------------------------
2019-07-17 13:42:16,312 Model training base path: "."
2019-07-17 13:42:16,314 ----------------------------------------------------------------------------------------------------
2019-07-17 13:42:16,316 Device: cuda:0
2019-07-17 13:42:16,318 ----------------------------------------------------------------------------------------------------
2019-07-17 13:42:16,320 Embedding storage mode: cpu
2019-07-17 13:42:16,322 ----------------------------------------------------------------------------------------------------
2019-07-17 13:42:16,980 epoch 1 - iter 0/3 - loss 1.12953818
2019-07-17 13:42:17,505 epoch 1 - iter 1/3 - loss 0.88430059
2019-07-17 13:42:17,838 epoch 1 - iter 2/3 - loss 0.68384539
2019-07-17 13:42:17,860 ----------------------------------------------------------------------------------------------------
2019-07-17 13:42:17,861 EPOCH 1 done: loss 0.6838 - lr 0.1000
2019-07-17 13:42:18,233 DEV : loss 0.2169550657272339 - score 0.95
2019-07-17 13:42:18,265 BAD EPOCHS (no improvement): 0
2019-07-17 13:42:29,995 ----------------------------------------------------------------------------------------------------
2019-07-17 13:42:30,097 epoch 2 - iter 0/3 - loss 0.24342805
2019-07-17 13:42:30,217 epoch 2 - iter 1/3 - loss 0.25471178
2019-07-17 13:42:30,266 epoch 2 - iter 2/3 - loss 0.23017298
2019-07-17 13:42:30,284 ----------------------------------------------------------------------------------------------------
2019-07-17 13:42:30,285 EPOCH 2 done: loss 0.2302 - lr 0.1000
2019-07-17 13:42:30,316 DEV : loss 0.1935195028781891 - score 0.95
2019-07-17 13:42:30,320 BAD EPOCHS (no improvement): 1
2019-07-17 13:42:41,949 ----------------------------------------------------------------------------------------------------
2019-07-17 13:42:42,050 epoch 3 - iter 0/3 - loss 0.11351790
2019-07-17 13:42:42,171 epoch 3 - iter 1/3 - loss 0.13642221
2019-07-17 13:42:42,239 epoch 3 - iter 2/3 - loss 0.16537253
2019-07-17 13:42:42,258 ----------------------------------------------------------------------------------------------------
2019-07-17 13:42:42,259 EPOCH 3 done: loss 0.1654 - lr 0.1000
2019-07-17 13:42:42,294 DEV : loss 0.19953027367591858 - score 0.95
2019-07-17 13:42:42,298 BAD EPOCHS (no improvement): 2
2019-07-17 13:42:53,960 ----------------------------------------------------------------------------------------------------
2019-07-17 13:42:54,060 epoch 4 - iter 0/3 - loss 0.14812917
2019-07-17 13:42:54,167 epoch 4 - iter 1/3 - loss 0.10075037
2019-07-17 13:42:54,234 epoch 4 - iter 2/3 - loss 0.11099806
2019-07-17 13:42:54,252 ----------------------------------------------------------------------------------------------------
2019-07-17 13:42:54,253 EPOCH 4 done: loss 0.1110 - lr 0.1000
2019-07-17 13:42:54,290 DEV : loss 0.17750518023967743 - score 0.95
2019-07-17 13:42:54,294 BAD EPOCHS (no improvement): 3
2019-07-17 13:43:05,926 ----------------------------------------------------------------------------------------------------
2019-07-17 13:43:06,024 epoch 5 - iter 0/3 - loss 0.07902250
2019-07-17 13:43:06,133 epoch 5 - iter 1/3 - loss 0.07272547
2019-07-17 13:43:06,225 epoch 5 - iter 2/3 - loss 0.07180052
2019-07-17 13:43:06,242 ----------------------------------------------------------------------------------------------------
2019-07-17 13:43:06,243 EPOCH 5 done: loss 0.0718 - lr 0.1000
2019-07-17 13:43:06,278 DEV : loss 0.18163713812828064 - score 0.95
Epoch 4: reducing learning rate of group 0 to 5.0000e-02.
2019-07-17 13:43:06,282 BAD EPOCHS (no improvement): 4
2019-07-17 13:43:18,202 ----------------------------------------------------------------------------------------------------
2019-07-17 13:43:18,407 epoch 6 - iter 0/3 - loss 0.04975884
2019-07-17 13:43:18,523 epoch 6 - iter 1/3 - loss 0.05130858
2019-07-17 13:43:18,589 epoch 6 - iter 2/3 - loss 0.05122599
2019-07-17 13:43:18,607 ----------------------------------------------------------------------------------------------------
2019-07-17 13:43:18,608 EPOCH 6 done: loss 0.0512 - lr 0.0500
2019-07-17 13:43:18,651 DEV : loss 0.17876218259334564 - score 0.95
2019-07-17 13:43:18,655 BAD EPOCHS (no improvement): 1
2019-07-17 13:43:30,066 ----------------------------------------------------------------------------------------------------
2019-07-17 13:43:30,184 epoch 7 - iter 0/3 - loss 0.04615138
2019-07-17 13:43:30,302 epoch 7 - iter 1/3 - loss 0.04906042
2019-07-17 13:43:30,365 epoch 7 - iter 2/3 - loss 0.03841856
2019-07-17 13:43:30,383 ----------------------------------------------------------------------------------------------------
2019-07-17 13:43:30,384 EPOCH 7 done: loss 0.0384 - lr 0.0500
2019-07-17 13:43:30,419 DEV : loss 0.1795329749584198 - score 0.95
2019-07-17 13:43:30,423 BAD EPOCHS (no improvement): 2
2019-07-17 13:43:42,185 ----------------------------------------------------------------------------------------------------
2019-07-17 13:43:42,445 epoch 8 - iter 0/3 - loss 0.04269283
2019-07-17 13:43:42,552 epoch 8 - iter 1/3 - loss 0.04601793
2019-07-17 13:43:42,613 epoch 8 - iter 2/3 - loss 0.03586188
2019-07-17 13:43:42,631 ----------------------------------------------------------------------------------------------------
2019-07-17 13:43:42,632 EPOCH 8 done: loss 0.0359 - lr 0.0500
2019-07-17 13:43:42,665 DEV : loss 0.18019260466098785 - score 0.95
2019-07-17 13:43:42,669 BAD EPOCHS (no improvement): 3
2019-07-17 13:43:54,199 ----------------------------------------------------------------------------------------------------
2019-07-17 13:43:54,508 epoch 9 - iter 0/3 - loss 0.04562064
2019-07-17 13:43:54,606 epoch 9 - iter 1/3 - loss 0.03992913
2019-07-17 13:43:54,666 epoch 9 - iter 2/3 - loss 0.03951348
2019-07-17 13:43:54,684 ----------------------------------------------------------------------------------------------------
2019-07-17 13:43:54,685 EPOCH 9 done: loss 0.0395 - lr 0.0500
2019-07-17 13:43:54,721 DEV : loss 0.17171242833137512 - score 0.95
Epoch 8: reducing learning rate of group 0 to 2.5000e-02.
2019-07-17 13:43:54,724 BAD EPOCHS (no improvement): 4
2019-07-17 13:44:06,064 ----------------------------------------------------------------------------------------------------
2019-07-17 13:44:06,357 epoch 10 - iter 0/3 - loss 0.02704864
2019-07-17 13:44:06,460 epoch 10 - iter 1/3 - loss 0.03585679
2019-07-17 13:44:06,517 epoch 10 - iter 2/3 - loss 0.03025564
2019-07-17 13:44:06,535 ----------------------------------------------------------------------------------------------------
2019-07-17 13:44:06,535 EPOCH 10 done: loss 0.0303 - lr 0.0250
2019-07-17 13:44:06,569 DEV : loss 0.1731186807155609 - score 0.95
2019-07-17 13:44:06,573 BAD EPOCHS (no improvement): 1
2019-07-17 13:44:18,198 ----------------------------------------------------------------------------------------------------
2019-07-17 13:44:18,430 epoch 11 - iter 0/3 - loss 0.02249379
2019-07-17 13:44:18,541 epoch 11 - iter 1/3 - loss 0.03357615
2019-07-17 13:44:18,605 epoch 11 - iter 2/3 - loss 0.02870297
2019-07-17 13:44:18,623 ----------------------------------------------------------------------------------------------------
2019-07-17 13:44:18,624 EPOCH 11 done: loss 0.0287 - lr 0.0250
2019-07-17 13:44:18,664 DEV : loss 0.17455197870731354 - score 0.95
2019-07-17 13:44:18,668 BAD EPOCHS (no improvement): 2
2019-07-17 13:44:30,044 ----------------------------------------------------------------------------------------------------
2019-07-17 13:44:30,351 epoch 12 - iter 0/3 - loss 0.01436380
2019-07-17 13:44:30,461 epoch 12 - iter 1/3 - loss 0.02262733
2019-07-17 13:44:30,533 epoch 12 - iter 2/3 - loss 0.03364742
2019-07-17 13:44:30,551 ----------------------------------------------------------------------------------------------------
2019-07-17 13:44:30,552 EPOCH 12 done: loss 0.0336 - lr 0.0250
2019-07-17 13:44:30,587 DEV : loss 0.16914425790309906 - score 0.95
2019-07-17 13:44:30,591 BAD EPOCHS (no improvement): 3
2019-07-17 13:44:42,223 ----------------------------------------------------------------------------------------------------
2019-07-17 13:44:42,484 epoch 13 - iter 0/3 - loss 0.02084849
2019-07-17 13:44:42,591 epoch 13 - iter 1/3 - loss 0.03388120
2019-07-17 13:44:42,653 epoch 13 - iter 2/3 - loss 0.02657653
2019-07-17 13:44:42,670 ----------------------------------------------------------------------------------------------------
2019-07-17 13:44:42,671 EPOCH 13 done: loss 0.0266 - lr 0.0250
2019-07-17 13:44:42,705 DEV : loss 0.1697494089603424 - score 0.95
Epoch 12: reducing learning rate of group 0 to 1.2500e-02.
2019-07-17 13:44:42,709 BAD EPOCHS (no improvement): 4
2019-07-17 13:44:54,050 ----------------------------------------------------------------------------------------------------
2019-07-17 13:44:54,304 epoch 14 - iter 0/3 - loss 0.02188246
2019-07-17 13:44:54,409 epoch 14 - iter 1/3 - loss 0.02183745
2019-07-17 13:44:54,464 epoch 14 - iter 2/3 - loss 0.02761706
2019-07-17 13:44:54,482 ----------------------------------------------------------------------------------------------------
2019-07-17 13:44:54,483 EPOCH 14 done: loss 0.0276 - lr 0.0125
2019-07-17 13:44:54,516 DEV : loss 0.17058534920215607 - score 0.95
2019-07-17 13:44:54,522 BAD EPOCHS (no improvement): 1
2019-07-17 13:45:06,341 ----------------------------------------------------------------------------------------------------
2019-07-17 13:45:06,576 epoch 15 - iter 0/3 - loss 0.01772359
2019-07-17 13:45:06,687 epoch 15 - iter 1/3 - loss 0.02583567
2019-07-17 13:45:06,745 epoch 15 - iter 2/3 - loss 0.02446777
2019-07-17 13:45:06,763 ----------------------------------------------------------------------------------------------------
2019-07-17 13:45:06,764 EPOCH 15 done: loss 0.0245 - lr 0.0125
2019-07-17 13:45:06,798 DEV : loss 0.17106768488883972 - score 0.95
2019-07-17 13:45:06,803 BAD EPOCHS (no improvement): 2
2019-07-17 13:45:18,339 ----------------------------------------------------------------------------------------------------
2019-07-17 13:45:18,662 epoch 16 - iter 0/3 - loss 0.02152755
2019-07-17 13:45:18,770 epoch 16 - iter 1/3 - loss 0.02693469
2019-07-17 13:45:18,834 epoch 16 - iter 2/3 - loss 0.02570675
2019-07-17 13:45:18,852 ----------------------------------------------------------------------------------------------------
2019-07-17 13:45:18,853 EPOCH 16 done: loss 0.0257 - lr 0.0125
2019-07-17 13:45:18,890 DEV : loss 0.17177173495292664 - score 0.95
2019-07-17 13:45:18,894 BAD EPOCHS (no improvement): 3
2019-07-17 13:45:30,717 ----------------------------------------------------------------------------------------------------
2019-07-17 13:45:30,988 epoch 17 - iter 0/3 - loss 0.04182786
2019-07-17 13:45:31,098 epoch 17 - iter 1/3 - loss 0.02848937
2019-07-17 13:45:31,165 epoch 17 - iter 2/3 - loss 0.02494488
2019-07-17 13:45:31,184 ----------------------------------------------------------------------------------------------------
2019-07-17 13:45:31,185 EPOCH 17 done: loss 0.0249 - lr 0.0125
2019-07-17 13:45:31,219 DEV : loss 0.17202387750148773 - score 0.95
Epoch 16: reducing learning rate of group 0 to 6.2500e-03.
2019-07-17 13:45:31,223 BAD EPOCHS (no improvement): 4
2019-07-17 13:45:43,053 ----------------------------------------------------------------------------------------------------
2019-07-17 13:45:43,286 epoch 18 - iter 0/3 - loss 0.03713746
2019-07-17 13:45:43,396 epoch 18 - iter 1/3 - loss 0.02629910
2019-07-17 13:45:43,457 epoch 18 - iter 2/3 - loss 0.02328312
2019-07-17 13:45:43,477 ----------------------------------------------------------------------------------------------------
2019-07-17 13:45:43,478 EPOCH 18 done: loss 0.0233 - lr 0.0063
2019-07-17 13:45:43,512 DEV : loss 0.17235131561756134 - score 0.95
2019-07-17 13:45:43,516 BAD EPOCHS (no improvement): 1
2019-07-17 13:45:55,936 ----------------------------------------------------------------------------------------------------
2019-07-17 13:45:56,038 epoch 19 - iter 0/3 - loss 0.01531454
2019-07-17 13:45:56,161 epoch 19 - iter 1/3 - loss 0.01697974
2019-07-17 13:45:56,232 epoch 19 - iter 2/3 - loss 0.02883327
2019-07-17 13:45:56,250 ----------------------------------------------------------------------------------------------------
2019-07-17 13:45:56,251 EPOCH 19 done: loss 0.0288 - lr 0.0063
2019-07-17 13:45:56,286 DEV : loss 0.17175742983818054 - score 0.95
2019-07-17 13:45:56,290 BAD EPOCHS (no improvement): 2
2019-07-17 13:46:08,008 ----------------------------------------------------------------------------------------------------
2019-07-17 13:46:08,203 epoch 20 - iter 0/3 - loss 0.01971393
2019-07-17 13:46:08,318 epoch 20 - iter 1/3 - loss 0.02337951
2019-07-17 13:46:08,377 epoch 20 - iter 2/3 - loss 0.02466225
2019-07-17 13:46:08,394 ----------------------------------------------------------------------------------------------------
2019-07-17 13:46:08,395 EPOCH 20 done: loss 0.0247 - lr 0.0063
2019-07-17 13:46:08,428 DEV : loss 0.17199192941188812 - score 0.95
2019-07-17 13:46:08,433 BAD EPOCHS (no improvement): 3
2019-07-17 13:46:20,037 ----------------------------------------------------------------------------------------------------
2019-07-17 13:46:20,353 epoch 21 - iter 0/3 - loss 0.03393368
2019-07-17 13:46:20,461 epoch 21 - iter 1/3 - loss 0.02603878
2019-07-17 13:46:20,534 epoch 21 - iter 2/3 - loss 0.02404766
2019-07-17 13:46:20,552 ----------------------------------------------------------------------------------------------------
2019-07-17 13:46:20,553 EPOCH 21 done: loss 0.0240 - lr 0.0063
2019-07-17 13:46:20,591 DEV : loss 0.17236179113388062 - score 0.95
Epoch 20: reducing learning rate of group 0 to 3.1250e-03.
2019-07-17 13:46:20,594 BAD EPOCHS (no improvement): 4
2019-07-17 13:46:32,294 ----------------------------------------------------------------------------------------------------
2019-07-17 13:46:32,513 epoch 22 - iter 0/3 - loss 0.02458208
2019-07-17 13:46:32,616 epoch 22 - iter 1/3 - loss 0.02049965
2019-07-17 13:46:32,673 epoch 22 - iter 2/3 - loss 0.02479812
2019-07-17 13:46:32,691 ----------------------------------------------------------------------------------------------------
2019-07-17 13:46:32,692 EPOCH 22 done: loss 0.0248 - lr 0.0031
2019-07-17 13:46:32,724 DEV : loss 0.17253009974956512 - score 0.95
2019-07-17 13:46:32,728 BAD EPOCHS (no improvement): 1
2019-07-17 13:46:44,341 ----------------------------------------------------------------------------------------------------
2019-07-17 13:46:44,633 epoch 23 - iter 0/3 - loss 0.03389432
2019-07-17 13:46:44,733 epoch 23 - iter 1/3 - loss 0.03052446
2019-07-17 13:46:44,798 epoch 23 - iter 2/3 - loss 0.02322454
2019-07-17 13:46:44,816 ----------------------------------------------------------------------------------------------------
2019-07-17 13:46:44,816 EPOCH 23 done: loss 0.0232 - lr 0.0031
2019-07-17 13:46:44,855 DEV : loss 0.17259268462657928 - score 0.95
2019-07-17 13:46:44,859 BAD EPOCHS (no improvement): 2
2019-07-17 13:46:56,977 ----------------------------------------------------------------------------------------------------
2019-07-17 13:46:57,181 epoch 24 - iter 0/3 - loss 0.01864637
2019-07-17 13:46:57,291 epoch 24 - iter 1/3 - loss 0.02823427
2019-07-17 13:46:57,346 epoch 24 - iter 2/3 - loss 0.02560668
2019-07-17 13:46:57,364 ----------------------------------------------------------------------------------------------------
2019-07-17 13:46:57,365 EPOCH 24 done: loss 0.0256 - lr 0.0031
2019-07-17 13:46:57,398 DEV : loss 0.1724378615617752 - score 0.95
2019-07-17 13:46:57,402 BAD EPOCHS (no improvement): 3
2019-07-17 13:47:09,119 ----------------------------------------------------------------------------------------------------
2019-07-17 13:47:09,374 epoch 25 - iter 0/3 - loss 0.00892743
2019-07-17 13:47:09,476 epoch 25 - iter 1/3 - loss 0.01906814
2019-07-17 13:47:09,548 epoch 25 - iter 2/3 - loss 0.02745630
2019-07-17 13:47:09,566 ----------------------------------------------------------------------------------------------------
2019-07-17 13:47:09,567 EPOCH 25 done: loss 0.0275 - lr 0.0031
2019-07-17 13:47:09,601 DEV : loss 0.17208600044250488 - score 0.95
Epoch 24: reducing learning rate of group 0 to 1.5625e-03.
2019-07-17 13:47:09,605 BAD EPOCHS (no improvement): 4
2019-07-17 13:47:20,979 ----------------------------------------------------------------------------------------------------
2019-07-17 13:47:21,289 epoch 26 - iter 0/3 - loss 0.02161277
2019-07-17 13:47:21,395 epoch 26 - iter 1/3 - loss 0.02249090
2019-07-17 13:47:21,454 epoch 26 - iter 2/3 - loss 0.02527559
2019-07-17 13:47:21,473 ----------------------------------------------------------------------------------------------------
2019-07-17 13:47:21,474 EPOCH 26 done: loss 0.0253 - lr 0.0016
2019-07-17 13:47:21,507 DEV : loss 0.1720239520072937 - score 0.95
2019-07-17 13:47:21,511 BAD EPOCHS (no improvement): 1
2019-07-17 13:47:33,272 ----------------------------------------------------------------------------------------------------
2019-07-17 13:47:33,497 epoch 27 - iter 0/3 - loss 0.02178361
2019-07-17 13:47:33,604 epoch 27 - iter 1/3 - loss 0.02418461
2019-07-17 13:47:33,667 epoch 27 - iter 2/3 - loss 0.02266802
2019-07-17 13:47:33,686 ----------------------------------------------------------------------------------------------------
2019-07-17 13:47:33,687 EPOCH 27 done: loss 0.0227 - lr 0.0016
2019-07-17 13:47:33,721 DEV : loss 0.1721399575471878 - score 0.95
2019-07-17 13:47:33,725 BAD EPOCHS (no improvement): 2
2019-07-17 13:47:45,251 ----------------------------------------------------------------------------------------------------
2019-07-17 13:47:45,541 epoch 28 - iter 0/3 - loss 0.03237828
2019-07-17 13:47:45,649 epoch 28 - iter 1/3 - loss 0.02844849
2019-07-17 13:47:45,719 epoch 28 - iter 2/3 - loss 0.02257865
2019-07-17 13:47:45,737 ----------------------------------------------------------------------------------------------------
2019-07-17 13:47:45,738 EPOCH 28 done: loss 0.0226 - lr 0.0016
2019-07-17 13:47:45,771 DEV : loss 0.17222264409065247 - score 0.95
2019-07-17 13:47:45,775 BAD EPOCHS (no improvement): 3
2019-07-17 13:47:57,394 ----------------------------------------------------------------------------------------------------
2019-07-17 13:47:57,650 epoch 29 - iter 0/3 - loss 0.01375472
2019-07-17 13:47:57,757 epoch 29 - iter 1/3 - loss 0.02176850
2019-07-17 13:47:57,817 epoch 29 - iter 2/3 - loss 0.02888534
2019-07-17 13:47:57,836 ----------------------------------------------------------------------------------------------------
2019-07-17 13:47:57,837 EPOCH 29 done: loss 0.0289 - lr 0.0016
2019-07-17 13:47:57,869 DEV : loss 0.17208896577358246 - score 0.95
Epoch 28: reducing learning rate of group 0 to 7.8125e-04.
2019-07-17 13:47:57,873 BAD EPOCHS (no improvement): 4
2019-07-17 13:48:09,462 ----------------------------------------------------------------------------------------------------
2019-07-17 13:48:09,773 epoch 30 - iter 0/3 - loss 0.01996330
2019-07-17 13:48:09,884 epoch 30 - iter 1/3 - loss 0.02312580
2019-07-17 13:48:09,951 epoch 30 - iter 2/3 - loss 0.02738177
2019-07-17 13:48:09,969 ----------------------------------------------------------------------------------------------------
2019-07-17 13:48:09,970 EPOCH 30 done: loss 0.0274 - lr 0.0008
2019-07-17 13:48:10,009 DEV : loss 0.17210647463798523 - score 0.95
2019-07-17 13:48:10,013 BAD EPOCHS (no improvement): 1
2019-07-17 13:48:27,824 ----------------------------------------------------------------------------------------------------
2019-07-17 13:48:28,038 Testing using best model ...
2019-07-17 13:48:28,044 loading file best-model.pt
2019-07-17 13:48:34,729 0.7588 0.7588 0.7588
2019-07-17 13:48:34,730
MICRO_AVG: acc 0.6113 - f1-score 0.7588
MACRO_AVG: acc 0.2529 - f1-score 0.28763333333333335
0 tp: 195 - fp: 62 - fn: 0 - tn: 0 - precision: 0.7588 - recall: 1.0000 - accuracy: 0.7588 - f1-score: 0.8629
1 tp: 0 - fp: 0 - fn: 44 - tn: 213 - precision: 0.0000 - recall: 0.0000 - accuracy: 0.0000 - f1-score: 0.0000
2 tp: 0 - fp: 0 - fn: 18 - tn: 239 - precision: 0.0000 - recall: 0.0000 - accuracy: 0.0000 - f1-score: 0.0000
2019-07-17 13:48:34,731 ----------------------------------------------------------------------------------------------------
Out[46]:
{'dev_loss_history': [tensor(0.2170, device='cuda:0'),
tensor(0.1935, device='cuda:0'),
tensor(0.1995, device='cuda:0'),
tensor(0.1775, device='cuda:0'),
tensor(0.1816, device='cuda:0'),
tensor(0.1788, device='cuda:0'),
tensor(0.1795, device='cuda:0'),
tensor(0.1802, device='cuda:0'),
tensor(0.1717, device='cuda:0'),
tensor(0.1731, device='cuda:0'),
tensor(0.1746, device='cuda:0'),
tensor(0.1691, device='cuda:0'),
tensor(0.1697, device='cuda:0'),
tensor(0.1706, device='cuda:0'),
tensor(0.1711, device='cuda:0'),
tensor(0.1718, device='cuda:0'),
tensor(0.1720, device='cuda:0'),
tensor(0.1724, device='cuda:0'),
tensor(0.1718, device='cuda:0'),
tensor(0.1720, device='cuda:0'),
tensor(0.1724, device='cuda:0'),
tensor(0.1725, device='cuda:0'),
tensor(0.1726, device='cuda:0'),
tensor(0.1724, device='cuda:0'),
tensor(0.1721, device='cuda:0'),
tensor(0.1720, device='cuda:0'),
tensor(0.1721, device='cuda:0'),
tensor(0.1722, device='cuda:0'),
tensor(0.1721, device='cuda:0'),
tensor(0.1721, device='cuda:0')],
'dev_score_history': [0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95,
0.95],
'test_score': 0.7588,
'train_loss_history': [0.6838453908761343,
0.23017298181851706,
0.16537252813577652,
0.11099805558721225,
0.07180052250623703,
0.05122598881522814,
0.038418558736642204,
0.03586187958717346,
0.0395134761929512,
0.030255640546480816,
0.02870296686887741,
0.033647422989209495,
0.026576528946558636,
0.02761705716451009,
0.024467766284942627,
0.02570674940943718,
0.02494487663110097,
0.023283122728268307,
0.02883326510588328,
0.024662251273790996,
0.024047664056221645,
0.0247981163362662,
0.023224542538324993,
0.025606680661439896,
0.027456300954023998,
0.02527559424440066,
0.02266802390416463,
0.022578647981087368,
0.028885335971911747,
0.027381767829259235]}
In [0]:
import os
import math
import random
# os.remove("trainkomm.txt")
# os.remove("testkomm.txt")
# os.remove("devkomm.txt")
listrand = list(range(100))
random.shuffle(listrand)
listranddev = []
for i in range(math.floor(len(listrand)*0.2)):
listranddev.append(listrand.pop())
file1 = open("trainkomm.txt","a", encoding="utf-8")
for i in listrand:
file1.writelines(['__label__',str(dfcred['stance komunikasi'].iloc[i]),' ',str(dfcred['sentence'].iloc[i]),'\n'])
file1.close()
file2 = open("devkomm.txt","a", encoding="utf-8")
for i in listranddev:
file2.writelines(['__label__',str(dfcred['stance komunikasi'].iloc[i]),' ',str(dfcred['sentence'].iloc[i]),'\n'])
file2.close()
file3 = open("testkomm.txt","a", encoding="utf-8")
for i in range(len(dftestcred)):
file3.writelines(['__label__',str(dftestcred['stance komunikasi'].iloc[i]),' ',str(dftestcred['sentence'].iloc[i]),'\n'])
file3.close()
In [48]:
from flair.data_fetcher import NLPTaskDataFetcher
from flair.embeddings import WordEmbeddings, FlairEmbeddings, DocumentRNNEmbeddings
from flair.models import TextClassifier
from flair.trainers import ModelTrainer
from pathlib import Path
corpus = NLPTaskDataFetcher.load_classification_corpus(Path('./'), test_file='testkomm.txt', dev_file='devkomm.txt', train_file='trainkomm.txt')
word_embeddings = [WordEmbeddings('glove'), FlairEmbeddings('id-forward'), FlairEmbeddings('id-backward')]
document_embeddings = DocumentRNNEmbeddings(word_embeddings, hidden_size=256, bidirectional=True,rnn_type='lstm',rnn_layers=1,dropout=0.2)
classifier = TextClassifier(document_embeddings, label_dictionary=corpus.make_label_dictionary(), multi_label=False)
trainer = ModelTrainer(classifier, corpus)
trainer.train('./', max_epochs=30,checkpoint=True)
!mv best-model.pt best-komm-model.pt
2019-07-17 13:49:22,888 Reading data from .
2019-07-17 13:49:22,890 Train: trainkomm.txt
2019-07-17 13:49:22,891 Dev: devkomm.txt
2019-07-17 13:49:22,893 Test: testkomm.txt
/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:7: DeprecationWarning: Call to deprecated function (or staticmethod) load_classification_corpus. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
import sys
/usr/local/lib/python3.6/dist-packages/flair/data_fetcher.py:447: DeprecationWarning: Call to deprecated function (or staticmethod) read_text_classification_file. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
max_tokens_per_doc=max_tokens_per_doc,
/usr/local/lib/python3.6/dist-packages/flair/data_fetcher.py:454: DeprecationWarning: Call to deprecated function (or staticmethod) read_text_classification_file. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
max_tokens_per_doc=max_tokens_per_doc,
/usr/local/lib/python3.6/dist-packages/flair/data_fetcher.py:463: DeprecationWarning: Call to deprecated function (or staticmethod) read_text_classification_file. (Use 'flair.datasets' instead.) -- Deprecated since version 0.4.1.
max_tokens_per_doc=max_tokens_per_doc,
/usr/local/lib/python3.6/dist-packages/smart_open/smart_open_lib.py:398: UserWarning: This function is deprecated, use smart_open.open instead. See the migration notes for details: https://github.com/RaRe-Technologies/smart_open/blob/master/README.rst#migrating-to-the-new-open-function
'See the migration notes for details: %s' % _MIGRATION_NOTES_URL
2019-07-17 13:49:25,630 Computing label dictionary. Progress:
100%|██████████| 80/80 [00:00<00:00, 31167.04it/s]
2019-07-17 13:49:25,640 [b'2', b'0', b'3', b'1']
2019-07-17 13:49:25,808 ----------------------------------------------------------------------------------------------------
2019-07-17 13:49:25,809 Model: "TextClassifier(
(document_embeddings): DocumentRNNEmbeddings(
(embeddings): StackedEmbeddings(
(list_embedding_0): WordEmbeddings('glove')
(list_embedding_1): FlairEmbeddings(
(lm): LanguageModel(
(drop): Dropout(p=0.1)
(encoder): Embedding(8823, 100)
(rnn): LSTM(100, 2048)
(decoder): Linear(in_features=2048, out_features=8823, bias=True)
)
)
(list_embedding_2): FlairEmbeddings(
(lm): LanguageModel(
(drop): Dropout(p=0.1)
(encoder): Embedding(8823, 100)
(rnn): LSTM(100, 2048)
(decoder): Linear(in_features=2048, out_features=8823, bias=True)
)
)
)
(word_reprojection_map): Linear(in_features=4196, out_features=4196, bias=True)
(rnn): GRU(4196, 256, bidirectional=True)
(dropout): Dropout(p=0.2)
)
(decoder): Linear(in_features=1024, out_features=4, bias=True)
(loss_function): CrossEntropyLoss()
)"
2019-07-17 13:49:25,811 ----------------------------------------------------------------------------------------------------
2019-07-17 13:49:25,813 Corpus: "Corpus: 80 train + 20 dev + 257 test sentences"
2019-07-17 13:49:25,815 ----------------------------------------------------------------------------------------------------
2019-07-17 13:49:25,816 Parameters:
2019-07-17 13:49:25,818 - learning_rate: "0.1"
2019-07-17 13:49:25,820 - mini_batch_size: "32"
2019-07-17 13:49:25,822 - patience: "3"
2019-07-17 13:49:25,823 - anneal_factor: "0.5"
2019-07-17 13:49:25,825 - max_epochs: "30"
2019-07-17 13:49:25,826 - shuffle: "True"
2019-07-17 13:49:25,828 - train_with_dev: "False"
2019-07-17 13:49:25,829 ----------------------------------------------------------------------------------------------------
2019-07-17 13:49:25,831 Model training base path: "."
2019-07-17 13:49:25,833 ----------------------------------------------------------------------------------------------------
2019-07-17 13:49:25,835 Device: cuda:0
2019-07-17 13:49:25,836 ----------------------------------------------------------------------------------------------------
2019-07-17 13:49:25,837 Embedding storage mode: cpu
2019-07-17 13:49:25,840 ----------------------------------------------------------------------------------------------------
2019-07-17 13:49:26,482 epoch 1 - iter 0/3 - loss 1.39324415
2019-07-17 13:49:27,066 epoch 1 - iter 1/3 - loss 1.33005446
2019-07-17 13:49:27,458 epoch 1 - iter 2/3 - loss 1.25697132
2019-07-17 13:49:27,476 ----------------------------------------------------------------------------------------------------
2019-07-17 13:49:27,477 EPOCH 1 done: loss 1.2570 - lr 0.1000
2019-07-17 13:49:27,755 DEV : loss 0.8324764370918274 - score 0.85
2019-07-17 13:49:27,783 BAD EPOCHS (no improvement): 0
2019-07-17 13:49:39,777 ----------------------------------------------------------------------------------------------------
2019-07-17 13:49:39,884 epoch 2 - iter 0/3 - loss 0.91782272
2019-07-17 13:49:40,038 epoch 2 - iter 1/3 - loss 0.76920408
2019-07-17 13:49:40,102 epoch 2 - iter 2/3 - loss 0.72096499
2019-07-17 13:49:40,120 ----------------------------------------------------------------------------------------------------
2019-07-17 13:49:40,121 EPOCH 2 done: loss 0.7210 - lr 0.1000
2019-07-17 13:49:40,152 DEV : loss 0.6134657859802246 - score 0.85
2019-07-17 13:49:40,156 BAD EPOCHS (no improvement): 1
2019-07-17 13:49:51,921 ----------------------------------------------------------------------------------------------------
2019-07-17 13:49:52,020 epoch 3 - iter 0/3 - loss 0.70540977
2019-07-17 13:49:52,176 epoch 3 - iter 1/3 - loss 0.61578450
2019-07-17 13:49:52,251 epoch 3 - iter 2/3 - loss 0.50192838
2019-07-17 13:49:52,269 ----------------------------------------------------------------------------------------------------
2019-07-17 13:49:52,270 EPOCH 3 done: loss 0.5019 - lr 0.1000
2019-07-17 13:49:52,302 DEV : loss 0.6987583041191101 - score 0.85
2019-07-17 13:49:52,305 BAD EPOCHS (no improvement): 2
2019-07-17 13:50:04,058 ----------------------------------------------------------------------------------------------------
2019-07-17 13:50:04,175 epoch 4 - iter 0/3 - loss 0.30166784
2019-07-17 13:50:04,328 epoch 4 - iter 1/3 - loss 0.31333694
2019-07-17 13:50:04,391 epoch 4 - iter 2/3 - loss 0.31866241
2019-07-17 13:50:04,410 ----------------------------------------------------------------------------------------------------
2019-07-17 13:50:04,411 EPOCH 4 done: loss 0.3187 - lr 0.1000
2019-07-17 13:50:04,443 DEV : loss 0.8797782063484192 - score 0.75
2019-07-17 13:50:04,447 BAD EPOCHS (no improvement): 3
2019-07-17 13:50:10,245 ----------------------------------------------------------------------------------------------------
2019-07-17 13:50:10,354 epoch 5 - iter 0/3 - loss 0.23756982
2019-07-17 13:50:10,484 epoch 5 - iter 1/3 - loss 0.30597646
2019-07-17 13:50:10,554 epoch 5 - iter 2/3 - loss 0.28119594
2019-07-17 13:50:10,574 ----------------------------------------------------------------------------------------------------
2019-07-17 13:50:10,575 EPOCH 5 done: loss 0.2812 - lr 0.1000
2019-07-17 13:50:10,606 DEV : loss 0.7777904868125916 - score 0.85
Epoch 4: reducing learning rate of group 0 to 5.0000e-02.
2019-07-17 13:50:10,611 BAD EPOCHS (no improvement): 4
2019-07-17 13:50:22,195 ----------------------------------------------------------------------------------------------------
2019-07-17 13:50:22,516 epoch 6 - iter 0/3 - loss 0.13137025
2019-07-17 13:50:22,626 epoch 6 - iter 1/3 - loss 0.17685094
2019-07-17 13:50:22,695 epoch 6 - iter 2/3 - loss 0.18590667
2019-07-17 13:50:22,722 ----------------------------------------------------------------------------------------------------
2019-07-17 13:50:22,723 EPOCH 6 done: loss 0.1859 - lr 0.0500
2019-07-17 13:50:22,753 DEV : loss 0.8207136392593384 - score 0.75
2019-07-17 13:50:22,756 BAD EPOCHS (no improvement): 1
2019-07-17 13:50:28,433 ----------------------------------------------------------------------------------------------------
2019-07-17 13:50:28,540 epoch 7 - iter 0/3 - loss 0.16320810
2019-07-17 13:50:28,649 epoch 7 - iter 1/3 - loss 0.16559741
2019-07-17 13:50:28,705 epoch 7 - iter 2/3 - loss 0.15589692
2019-07-17 13:50:28,723 ----------------------------------------------------------------------------------------------------
2019-07-17 13:50:28,724 EPOCH 7 done: loss 0.1559 - lr 0.0500
2019-07-17 13:50:28,754 DEV : loss 0.8023449182510376 - score 0.85
2019-07-17 13:50:28,757 BAD EPOCHS (no improvement): 2
2019-07-17 13:50:40,082 ----------------------------------------------------------------------------------------------------
2019-07-17 13:50:40,392 epoch 8 - iter 0/3 - loss 0.12668729
2019-07-17 13:50:40,494 epoch 8 - iter 1/3 - loss 0.13779722
2019-07-17 13:50:40,558 epoch 8 - iter 2/3 - loss 0.13232738
2019-07-17 13:50:40,575 ----------------------------------------------------------------------------------------------------
2019-07-17 13:50:40,576 EPOCH 8 done: loss 0.1323 - lr 0.0500
2019-07-17 13:50:40,607 DEV : loss 0.8198299407958984 - score 0.85
2019-07-17 13:50:40,610 BAD EPOCHS (no improvement): 3
2019-07-17 13:50:52,077 ----------------------------------------------------------------------------------------------------
2019-07-17 13:50:52,359 epoch 9 - iter 0/3 - loss 0.14204857
2019-07-17 13:50:52,478 epoch 9 - iter 1/3 - loss 0.13364631
2019-07-17 13:50:52,541 epoch 9 - iter 2/3 - loss 0.12469518
2019-07-17 13:50:52,562 ----------------------------------------------------------------------------------------------------
2019-07-17 13:50:52,563 EPOCH 9 done: loss 0.1247 - lr 0.0500
2019-07-17 13:50:52,593 DEV : loss 0.8408986926078796 - score 0.75
Epoch 8: reducing learning rate of group 0 to 2.5000e-02.
2019-07-17 13:50:52,596 BAD EPOCHS (no improvement): 4
2019-07-17 13:50:58,268 ----------------------------------------------------------------------------------------------------
2019-07-17 13:50:58,376 epoch 10 - iter 0/3 - loss 0.10978664
2019-07-17 13:50:58,526 epoch 10 - iter 1/3 - loss 0.10951795
2019-07-17 13:50:58,591 epoch 10 - iter 2/3 - loss 0.11940708
2019-07-17 13:50:58,610 ----------------------------------------------------------------------------------------------------
2019-07-17 13:50:58,610 EPOCH 10 done: loss 0.1194 - lr 0.0250
2019-07-17 13:50:58,645 DEV : loss 0.8397005200386047 - score 0.75
2019-07-17 13:50:58,649 BAD EPOCHS (no improvement): 1
2019-07-17 13:51:04,310 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:04,408 epoch 11 - iter 0/3 - loss 0.11941489
2019-07-17 13:51:04,546 epoch 11 - iter 1/3 - loss 0.11774428
2019-07-17 13:51:04,613 epoch 11 - iter 2/3 - loss 0.10719929
2019-07-17 13:51:04,632 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:04,633 EPOCH 11 done: loss 0.1072 - lr 0.0250
2019-07-17 13:51:04,667 DEV : loss 0.8473146557807922 - score 0.75
2019-07-17 13:51:04,670 BAD EPOCHS (no improvement): 2
2019-07-17 13:51:10,402 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:10,517 epoch 12 - iter 0/3 - loss 0.09594084
2019-07-17 13:51:10,653 epoch 12 - iter 1/3 - loss 0.10673052
2019-07-17 13:51:10,721 epoch 12 - iter 2/3 - loss 0.10339756
2019-07-17 13:51:10,739 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:10,740 EPOCH 12 done: loss 0.1034 - lr 0.0250
2019-07-17 13:51:10,773 DEV : loss 0.8533797264099121 - score 0.75
2019-07-17 13:51:10,777 BAD EPOCHS (no improvement): 3
2019-07-17 13:51:16,424 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:16,526 epoch 13 - iter 0/3 - loss 0.10790632
2019-07-17 13:51:16,653 epoch 13 - iter 1/3 - loss 0.11221693
2019-07-17 13:51:16,721 epoch 13 - iter 2/3 - loss 0.09658618
2019-07-17 13:51:16,740 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:16,741 EPOCH 13 done: loss 0.0966 - lr 0.0250
2019-07-17 13:51:16,770 DEV : loss 0.860796332359314 - score 0.75
Epoch 12: reducing learning rate of group 0 to 1.2500e-02.
2019-07-17 13:51:16,774 BAD EPOCHS (no improvement): 4
2019-07-17 13:51:22,570 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:22,682 epoch 14 - iter 0/3 - loss 0.09574795
2019-07-17 13:51:22,821 epoch 14 - iter 1/3 - loss 0.10391901
2019-07-17 13:51:22,886 epoch 14 - iter 2/3 - loss 0.08785142
2019-07-17 13:51:22,905 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:22,906 EPOCH 14 done: loss 0.0879 - lr 0.0125
2019-07-17 13:51:22,939 DEV : loss 0.8620648384094238 - score 0.75
2019-07-17 13:51:22,942 BAD EPOCHS (no improvement): 1
2019-07-17 13:51:28,602 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:28,713 epoch 15 - iter 0/3 - loss 0.08273102
2019-07-17 13:51:28,827 epoch 15 - iter 1/3 - loss 0.08748479
2019-07-17 13:51:28,903 epoch 15 - iter 2/3 - loss 0.09941445
2019-07-17 13:51:28,921 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:28,922 EPOCH 15 done: loss 0.0994 - lr 0.0125
2019-07-17 13:51:28,953 DEV : loss 0.8673663139343262 - score 0.75
2019-07-17 13:51:28,957 BAD EPOCHS (no improvement): 2
2019-07-17 13:51:35,162 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:35,253 epoch 16 - iter 0/3 - loss 0.08632563
2019-07-17 13:51:35,370 epoch 16 - iter 1/3 - loss 0.09184272
2019-07-17 13:51:35,443 epoch 16 - iter 2/3 - loss 0.09036340
2019-07-17 13:51:35,461 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:35,462 EPOCH 16 done: loss 0.0904 - lr 0.0125
2019-07-17 13:51:35,496 DEV : loss 0.8691703677177429 - score 0.75
2019-07-17 13:51:35,499 BAD EPOCHS (no improvement): 3
2019-07-17 13:51:41,234 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:41,337 epoch 17 - iter 0/3 - loss 0.08905771
2019-07-17 13:51:41,455 epoch 17 - iter 1/3 - loss 0.08230402
2019-07-17 13:51:41,529 epoch 17 - iter 2/3 - loss 0.09238478
2019-07-17 13:51:41,548 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:41,549 EPOCH 17 done: loss 0.0924 - lr 0.0125
2019-07-17 13:51:41,580 DEV : loss 0.8709818720817566 - score 0.75
Epoch 16: reducing learning rate of group 0 to 6.2500e-03.
2019-07-17 13:51:41,585 BAD EPOCHS (no improvement): 4
2019-07-17 13:51:47,280 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:47,390 epoch 18 - iter 0/3 - loss 0.08214936
2019-07-17 13:51:47,522 epoch 18 - iter 1/3 - loss 0.08817354
2019-07-17 13:51:47,592 epoch 18 - iter 2/3 - loss 0.08351414
2019-07-17 13:51:47,611 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:47,612 EPOCH 18 done: loss 0.0835 - lr 0.0063
2019-07-17 13:51:47,645 DEV : loss 0.8716562390327454 - score 0.75
2019-07-17 13:51:47,649 BAD EPOCHS (no improvement): 1
2019-07-17 13:51:53,488 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:53,593 epoch 19 - iter 0/3 - loss 0.07063945
2019-07-17 13:51:53,714 epoch 19 - iter 1/3 - loss 0.08429876
2019-07-17 13:51:53,788 epoch 19 - iter 2/3 - loss 0.08661981
2019-07-17 13:51:53,807 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:53,808 EPOCH 19 done: loss 0.0866 - lr 0.0063
2019-07-17 13:51:53,839 DEV : loss 0.871860146522522 - score 0.75
2019-07-17 13:51:53,842 BAD EPOCHS (no improvement): 2
2019-07-17 13:51:59,550 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:59,651 epoch 20 - iter 0/3 - loss 0.07559578
2019-07-17 13:51:59,769 epoch 20 - iter 1/3 - loss 0.09350671
2019-07-17 13:51:59,844 epoch 20 - iter 2/3 - loss 0.08353805
2019-07-17 13:51:59,862 ----------------------------------------------------------------------------------------------------
2019-07-17 13:51:59,862 EPOCH 20 done: loss 0.0835 - lr 0.0063
2019-07-17 13:51:59,894 DEV : loss 0.8733702898025513 - score 0.75
2019-07-17 13:51:59,899 BAD EPOCHS (no improvement): 3
2019-07-17 13:52:05,489 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:05,585 epoch 21 - iter 0/3 - loss 0.07174131
2019-07-17 13:52:05,715 epoch 21 - iter 1/3 - loss 0.08108623
2019-07-17 13:52:05,788 epoch 21 - iter 2/3 - loss 0.08490493
2019-07-17 13:52:05,806 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:05,807 EPOCH 21 done: loss 0.0849 - lr 0.0063
2019-07-17 13:52:05,841 DEV : loss 0.8737885355949402 - score 0.75
Epoch 20: reducing learning rate of group 0 to 3.1250e-03.
2019-07-17 13:52:05,845 BAD EPOCHS (no improvement): 4
2019-07-17 13:52:11,545 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:11,912 epoch 22 - iter 0/3 - loss 0.08405107
2019-07-17 13:52:12,021 epoch 22 - iter 1/3 - loss 0.08264503
2019-07-17 13:52:12,083 epoch 22 - iter 2/3 - loss 0.07504124
2019-07-17 13:52:12,101 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:12,101 EPOCH 22 done: loss 0.0750 - lr 0.0031
2019-07-17 13:52:12,132 DEV : loss 0.8740658760070801 - score 0.75
2019-07-17 13:52:12,136 BAD EPOCHS (no improvement): 1
2019-07-17 13:52:17,812 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:17,918 epoch 23 - iter 0/3 - loss 0.07463364
2019-07-17 13:52:18,045 epoch 23 - iter 1/3 - loss 0.08249114
2019-07-17 13:52:18,109 epoch 23 - iter 2/3 - loss 0.07939351
2019-07-17 13:52:18,128 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:18,129 EPOCH 23 done: loss 0.0794 - lr 0.0031
2019-07-17 13:52:18,160 DEV : loss 0.8741914629936218 - score 0.75
2019-07-17 13:52:18,163 BAD EPOCHS (no improvement): 2
2019-07-17 13:52:23,963 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:24,052 epoch 24 - iter 0/3 - loss 0.07170549
2019-07-17 13:52:24,174 epoch 24 - iter 1/3 - loss 0.08005657
2019-07-17 13:52:24,251 epoch 24 - iter 2/3 - loss 0.09082525
2019-07-17 13:52:24,268 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:24,269 EPOCH 24 done: loss 0.0908 - lr 0.0031
2019-07-17 13:52:24,303 DEV : loss 0.8759194612503052 - score 0.75
2019-07-17 13:52:24,306 BAD EPOCHS (no improvement): 3
2019-07-17 13:52:30,208 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:30,506 epoch 25 - iter 0/3 - loss 0.08777730
2019-07-17 13:52:30,614 epoch 25 - iter 1/3 - loss 0.08389030
2019-07-17 13:52:30,682 epoch 25 - iter 2/3 - loss 0.08022731
2019-07-17 13:52:30,700 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:30,701 EPOCH 25 done: loss 0.0802 - lr 0.0031
2019-07-17 13:52:30,736 DEV : loss 0.8766950368881226 - score 0.75
Epoch 24: reducing learning rate of group 0 to 1.5625e-03.
2019-07-17 13:52:30,740 BAD EPOCHS (no improvement): 4
2019-07-17 13:52:36,422 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:36,511 epoch 26 - iter 0/3 - loss 0.08921727
2019-07-17 13:52:36,641 epoch 26 - iter 1/3 - loss 0.08169654
2019-07-17 13:52:36,710 epoch 26 - iter 2/3 - loss 0.08780979
2019-07-17 13:52:36,729 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:36,730 EPOCH 26 done: loss 0.0878 - lr 0.0016
2019-07-17 13:52:36,764 DEV : loss 0.8770910501480103 - score 0.75
2019-07-17 13:52:36,767 BAD EPOCHS (no improvement): 1
2019-07-17 13:52:42,518 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:42,611 epoch 27 - iter 0/3 - loss 0.09658294
2019-07-17 13:52:42,737 epoch 27 - iter 1/3 - loss 0.08722028
2019-07-17 13:52:42,806 epoch 27 - iter 2/3 - loss 0.07782392
2019-07-17 13:52:42,824 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:42,825 EPOCH 27 done: loss 0.0778 - lr 0.0016
2019-07-17 13:52:42,855 DEV : loss 0.8774480819702148 - score 0.75
2019-07-17 13:52:42,858 BAD EPOCHS (no improvement): 2
2019-07-17 13:52:48,498 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:48,772 epoch 28 - iter 0/3 - loss 0.09627122
2019-07-17 13:52:48,884 epoch 28 - iter 1/3 - loss 0.09573792
2019-07-17 13:52:48,946 epoch 28 - iter 2/3 - loss 0.08189793
2019-07-17 13:52:48,963 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:48,964 EPOCH 28 done: loss 0.0819 - lr 0.0016
2019-07-17 13:52:48,996 DEV : loss 0.8779863119125366 - score 0.75
2019-07-17 13:52:48,999 BAD EPOCHS (no improvement): 3
2019-07-17 13:52:55,366 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:55,474 epoch 29 - iter 0/3 - loss 0.07442470
2019-07-17 13:52:55,586 epoch 29 - iter 1/3 - loss 0.08437029
2019-07-17 13:52:55,659 epoch 29 - iter 2/3 - loss 0.08150101
2019-07-17 13:52:55,678 ----------------------------------------------------------------------------------------------------
2019-07-17 13:52:55,679 EPOCH 29 done: loss 0.0815 - lr 0.0016
2019-07-17 13:52:55,710 DEV : loss 0.8782764673233032 - score 0.75
Epoch 28: reducing learning rate of group 0 to 7.8125e-04.
2019-07-17 13:52:55,715 BAD EPOCHS (no improvement): 4
2019-07-17 13:53:01,317 ----------------------------------------------------------------------------------------------------
2019-07-17 13:53:01,643 epoch 30 - iter 0/3 - loss 0.07396122
2019-07-17 13:53:01,750 epoch 30 - iter 1/3 - loss 0.07779591
2019-07-17 13:53:01,816 epoch 30 - iter 2/3 - loss 0.08388884
2019-07-17 13:53:01,834 ----------------------------------------------------------------------------------------------------
2019-07-17 13:53:01,834 EPOCH 30 done: loss 0.0839 - lr 0.0008
2019-07-17 13:53:01,868 DEV : loss 0.878522515296936 - score 0.75
2019-07-17 13:53:01,872 BAD EPOCHS (no improvement): 1
2019-07-17 13:53:13,278 ----------------------------------------------------------------------------------------------------
2019-07-17 13:53:13,502 Testing using best model ...
2019-07-17 13:53:13,504 loading file best-model.pt
2019-07-17 13:53:23,882 0.7237 0.7237 0.7237
2019-07-17 13:53:23,884
MICRO_AVG: acc 0.5671 - f1-score 0.7237
MACRO_AVG: acc 0.1973 - f1-score 0.2363
0 tp: 184 - fp: 58 - fn: 7 - tn: 8 - precision: 0.7603 - recall: 0.9634 - accuracy: 0.7390 - f1-score: 0.8499
1 tp: 0 - fp: 0 - fn: 19 - tn: 238 - precision: 0.0000 - recall: 0.0000 - accuracy: 0.0000 - f1-score: 0.0000
2 tp: 0 - fp: 0 - fn: 20 - tn: 237 - precision: 0.0000 - recall: 0.0000 - accuracy: 0.0000 - f1-score: 0.0000
3 tp: 2 - fp: 13 - fn: 25 - tn: 217 - precision: 0.1333 - recall: 0.0741 - accuracy: 0.0500 - f1-score: 0.0953
2019-07-17 13:53:23,884 ----------------------------------------------------------------------------------------------------
Out[48]:
{'dev_loss_history': [tensor(0.8325, device='cuda:0'),
tensor(0.6135, device='cuda:0'),
tensor(0.6988, device='cuda:0'),
tensor(0.8798, device='cuda:0'),
tensor(0.7778, device='cuda:0'),
tensor(0.8207, device='cuda:0'),
tensor(0.8023, device='cuda:0'),
tensor(0.8198, device='cuda:0'),
tensor(0.8409, device='cuda:0'),
tensor(0.8397, device='cuda:0'),
tensor(0.8473, device='cuda:0'),
tensor(0.8534, device='cuda:0'),
tensor(0.8608, device='cuda:0'),
tensor(0.8621, device='cuda:0'),
tensor(0.8674, device='cuda:0'),
tensor(0.8692, device='cuda:0'),
tensor(0.8710, device='cuda:0'),
tensor(0.8717, device='cuda:0'),
tensor(0.8719, device='cuda:0'),
tensor(0.8734, device='cuda:0'),
tensor(0.8738, device='cuda:0'),
tensor(0.8741, device='cuda:0'),
tensor(0.8742, device='cuda:0'),
tensor(0.8759, device='cuda:0'),
tensor(0.8767, device='cuda:0'),
tensor(0.8771, device='cuda:0'),
tensor(0.8774, device='cuda:0'),
tensor(0.8780, device='cuda:0'),
tensor(0.8783, device='cuda:0'),
tensor(0.8785, device='cuda:0')],
'dev_score_history': [0.85,
0.85,
0.85,
0.75,
0.85,
0.75,
0.85,
0.85,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75,
0.75],
'test_score': 0.7237,
'train_loss_history': [1.2569713195164998,
0.7209649880727133,
0.5019283791383108,
0.3186624050140381,
0.28119594355424243,
0.1859066734711329,
0.15589692195256552,
0.13232738027969995,
0.12469517687956493,
0.11940707514683406,
0.1071992889046669,
0.10339755813280742,
0.09658617526292801,
0.08785142004489899,
0.09941445291042328,
0.09036339819431305,
0.0923847829302152,
0.08351414402325948,
0.08661980926990509,
0.08353804796934128,
0.08490493148565292,
0.07504124194383621,
0.07939350605010986,
0.09082524975140889,
0.08022731045881908,
0.08780979365110397,
0.07782391955455144,
0.08189792931079865,
0.08150101453065872,
0.08388884365558624]}
In [49]:
uploaded = drive.CreateFile({'title': 'best-cred-model.pt'})
uploaded.SetContentFile('best-cred-model.pt')
uploaded.Upload()
print('Uploaded file with ID {}'.format(uploaded.get('id')))
uploaded = drive.CreateFile({'title': 'best-form-model.pt'})
uploaded.SetContentFile('best-form-model.pt')
uploaded.Upload()
print('Uploaded file with ID {}'.format(uploaded.get('id')))
uploaded = drive.CreateFile({'title': 'best-form-model.pt'})
uploaded.SetContentFile('best-eff-model.pt')
uploaded.Upload()
print('Uploaded file with ID {}'.format(uploaded.get('id')))
uploaded = drive.CreateFile({'title': 'best-eff-model.pt'})
adc.json devkomm.txt testeff.txt traineff.txt
best-model.pt devkoor.txt testform.txt trainform.txt
checkpoint.pt dev.txt testkomm.txt training.log
corpus_rev_cleaned_sample.xlsx final-model.pt testkoor.txt trainkomm.txt
corpus_rev_cleaned.xlsx loss.tsv test_sample.csv trainkoor.txt
deveff.txt sample_data test.tsv train.txt
devform.txt stopwordsid.txt test.txt weights.txt
Content source: TheJabbar/JamAPP
Similar notebooks: