Zambian Language Trainer Window
Curate local language datasets and trigger asynchronous machine learning training
Multilingual
Select Target Zambian Dialect to Manage & Train:
πΏπ²
Unified (All)
πΏπ²
Chibemba
πΏπ²
Chinyanja
πΏπ²
Chitonga
πΏπ²
Silozi
π¬π§
English
π
Mixed/General
Local Dataset Inspector: All
25506 Records| Phrase Text | Dialect | Category Label | Source |
|---|---|---|---|
| Nothing worse than a baby bitch pussy. Go fly a kite | Mixed | Abusive | User Added |
| Nothing says teabagger like a millionaire complaining that they are broke, huh #HillaryClinton? #Hillary2016, give me a break, lol. #p2 | Mixed | Safe | User Added |
| Nothing more yokel MSU fan than to constantly reassess Izzo *aka our only consistent winner since D. Daugherty or Biggie flipping Munn* | Mixed | Safe | User Added |
| Nothing is too spicy. And if it is, 1) stop being a fucking bitch 2) eat more of it 3) cry you fucking pussy | Mixed | Abusive | User Added |
| Nothing good on TV and the @Yankees are eating their balls 3 innings in. Great. | Mixed | Safe | User Added |
| Note: Khorasan is the old Arabic name for parts of Iran, Afghanistan. Not an invention of modern Jihadis. #justsayin | Mixed | Safe | User Added |
| Note the rocketeer removes cardboard from launcher and carries it to a trash bin before fleeing. #TidyTerrorist http://t.co/sb3U1XpeG4 | Mixed | Safe | User Added |
| Not ya grandma brownies, wink wink http://t.co/pZPI8RUQ4F | Mixed | Safe | User Added |
| Not trying to follow any spooks, no offense ... | Mixed | Hate Speech | User Added |
| Not true RT @_Kimchanel_: Every man has hoes #bye | Mixed | Abusive | User Added |
| Not trash, just plain as fuck for $225. And come on, no elephant print on a 3? | Mixed | Abusive | User Added |
| Not to sound like a total bitch but I'm tired of peoples kids and the back to school pics😒 we ain't post pics on the back to college day | Mixed | Abusive | User Added |
Add Local Phrase
Batch Import CSV
CSV OnlyUpload a custom CSV file to bulk import training data across all languages. Required columns: phrase (or text), category (0=Hate, 1=Abusive, 2=Safe), and optionally language.
Model Training Control
20%
Evaluation Results
Trained Algorithm:
Decision Tree
ACCURACY
F1 SCORE
Confusion Matrix Grid
46
0
0
2
41
4
0
0
47
Columns: Predicted
Rows: Actual
Algorithm Accuracy Graph