Views
No views yet
pip install transformers torch1from transformers import T5Tokenizer, T5ForConditionalGeneration
2
3# Model ve tokenizer yükleme
4model_name = "seyfullah2/turkish-stemmer-t5"
5tokenizer = T5Tokenizer.from_pretrained(model_name)
6model = T5ForConditionalGeneration.from_pretrained(model_name)
7
8# Kök bulma fonksiyonu
9def find_stem(word):
10 input_text = f"kök bul: {word}"
11 inputs = tokenizer(input_text, return_tensors="pt", max_length=64, truncation=True)
12
13 outputs = model.generate(
14 **inputs,
15 max_length=64,
16 num_beams=4,
17 early_stopping=True,
18 no_repeat_ngram_size=2
19 )
20
21 root = tokenizer.decode(outputs[0], skip_special_tokens=True)
22 return root.strip()
23
24# Örnek kullanım
25print(find_stem("kitaplardan")) # → kitap
26print(find_stem("evlerimizde")) # → ev
27print(find_stem("koşuyordum")) # → koş1words = ["kitaplar", "evlerde", "geliyorum", "başladı"]
2
3for word in words:
4 stem = find_stem(word)
5 print(f"{word} → {stem}")| Metrik | Değer |
|---|---|
| Accuracy | 92.00% |
| Test Samples | 5,500 |
| Avg Inference Time | ~50ms per word |
1@misc{turkish-stemmer-2025,
2 author = {seyfullah2},
3 title = {Turkish Stemmer T5},
4 year = {2025},
5 publisher = {Hugging Face},
6 howpublished = {\url{https://huggingface.co/seyfullah2/turkish-stemmer-t5}}
7}