Name		Name	Last commit message	Last commit date
Latest commit History 32 Commits
SSCL_Vocab500_RandomWeightedSampling		SSCL_Vocab500_RandomWeightedSampling
TrainingResult		TrainingResult
TrainingResults		TrainingResults
All_Models.py		All_Models.py
Constants.py		Constants.py
Datasets.py		Datasets.py
GatedCNN.png		GatedCNN.png
GatedCNN_Vocab500_RandomWeightedSampling.html		GatedCNN_Vocab500_RandomWeightedSampling.html
GatedCNN_Vocab8000_RandomWeightedSampling_WithDropout.html		GatedCNN_Vocab8000_RandomWeightedSampling_WithDropout.html
GatedCNN_Vocab8000_RandomWeightedSampling_WithDropout_Honeypot.html		GatedCNN_Vocab8000_RandomWeightedSampling_WithDropout_Honeypot.html
GatedCNN_Vocab8000_RandomWeightedSampling_WithDropout_lessSteps.html		GatedCNN_Vocab8000_RandomWeightedSampling_WithDropout_lessSteps.html
GridTkinterSpamDetection.py		GridTkinterSpamDetection.py
LoadData.py		LoadData.py
ModelAfter.ipynb		ModelAfter.ipynb
README.md		README.md
RefactoringTest.ipynb		RefactoringTest.ipynb
SSCL.png		SSCL.png
SSCL_Vocab500_RandomWeightedSampling.html		SSCL_Vocab500_RandomWeightedSampling.html
ScreenShot.png		ScreenShot.png
SelfAttn.png		SelfAttn.png
SelfAttn_Vocab500_RandomWeightedSampling.html		SelfAttn_Vocab500_RandomWeightedSampling.html
SelfAttn_Vocab8000_RandomWeightedSampling.html		SelfAttn_Vocab8000_RandomWeightedSampling.html
SpamDetectionModel.ipynb		SpamDetectionModel.ipynb
SpamDetectionWithUserInfo.ipynb		SpamDetectionWithUserInfo.ipynb
SubModels.py		SubModels.py
TestingPad.ipynb		TestingPad.ipynb
TkinterSpamDetection (copy).ipynb		TkinterSpamDetection (copy).ipynb
TkinterSpamDetection.ipynb		TkinterSpamDetection.ipynb
TkinterSpamDetection.py		TkinterSpamDetection.py
TkinterSpamDetectionGridVersion.ipynb		TkinterSpamDetectionGridVersion.ipynb
TkinterTry.ipynb		TkinterTry.ipynb
Trainer.py		Trainer.py
Trainers.py		Trainers.py
test.py		test.py
testingPad.ipynb		testingPad.ipynb
tkinterTest.py		tkinterTest.py
twitterIcon.png		twitterIcon.png
utils.py		utils.py

Repository files navigation

IFN702SpamDetection

To solve the problem of gradient vanish:

I have changed:

out = self.rnn(out)[0][:, -1, :]

to

out = self.rnn(out)[0].sum(dim=1)

Best tunned results

SSCL

0.9384

GatedCNN

0.9476

SelfAttn (Transformer)

0.9345

ChihchengHsieh/IFN702SpamDetection

Folders and files

Latest commit

History

Repository files navigation

IFN702SpamDetection

Best tunned results

SSCL

GatedCNN

SelfAttn (Transformer)

About

Resources

Stars

Watchers

Forks

Languages