Model is Turkish bert-based.
Used StratifiedKFold(5) for validation.
results [0.806858621805241, 0.8912621359223301, 0.9440129449838188, 0.9750809061488673, 0.9851132686084142]

Mean-Precision: 0.9204655754937342

	Uygun İlan	Is Ilani Degil	Mustehcen	Cift Pozisyon
Precision	0.986	0.996	0.966	0.970
Recall	0.992	0.986	0.966	0.959
F1 Score	0.989	0.991	0.966	0.965
Accuracy : 0.975

Example

!IMPORTANT_HINT: The sentence given to pipe must not contain Turkish characters.

from transformers import AutoTokenizer, TextClassificationPipeline, AutoModelForSequenceClassification

tokenizer = AutoTokenizer.from_pretrained("nanelimon/bert-base-turkish-job-advertisement")
model = AutoModelForSequenceClassification.from_pretrained("nanelimon/bert-base-turkish-job-advertisement")
pipe = TextClassificationPipeline(model=model, tokenizer=tokenizer)


def set_sentence(sentence: str):
    result = sentence.lower().replace('ö', 'o').replace('ı', 'i').replace('ü', 'u').replace('ç', 'c').replace('ğ', 'g').replace('ş', 's')
    return result


print(pipe(set_sentence('Fiziği düzgün 17 yaş kızlar aranıyor')))

Result;

output: [{'label': 'Mustehcen', 'score': 0.9992677569389343}]

label= It shows which class the sent Turkish text belongs to according to the model.
score= It shows the compliance rate of the Turkish text sent to the label found.

Authors

Seyma SARIGIL: seymasargil@gmail.com
Murat KOKLU: mkoklu@selcuk.edu.tr
Click to review Master's thesis

License

apache-2.0

Free Software, Hell Yeah!