Downloads · 30 days
0
prova-berturk-qlora
eminkutlu/prova-berturk-qlora
prova-berturk-qlora is a machine learning model from eminkutlu. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Türkçe multi-label behavioral/persona classifier.
Access
Public
Updated Sep 5, 2026
Repo size
10.8 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors10.8 MB · 91%
At a glance
- Access
- Public
- Created
- Sep 5, 2026
- Updated
- Sep 5, 2026
- SHA
- fe1b25d5
Model card
From the Hugging Face model README
Prova BERTurk QLoRA
Türkçe multi-label behavioral/persona classifier.
- Base model: dbmdz/bert-base-turkish-cased
- Dataset: eminkutlu/prova-dataset
- Method: QLoRA, 4-bit NF4 quantization
- Labels: pii_disclosure, identity_verification, greeting, active_listening, procedure_accuracy, empathy, solution_offer, closing
- Validation macro-F1: 0.9747
- Test macro-F1: 0.9821
Bu repo adapter ağırlıklarını içerir. Inference için base model ve PEFT adapter birlikte yüklenmelidir.
Örnek yükleme: from transformers import AutoTokenizer, AutoModelForSequenceClassification from peft import PeftModel base_id = "dbmdz/bert-base-turkish-cased" adapter_id = "eminkutlu/prova-berturk-qlora" tokenizer = AutoTokenizer.from_pretrained(adapter_id) base = AutoModelForSequenceClassification.from_pretrained(base_id, num_labels=8, problem_type="multi_label_classification") model = PeftModel.from_pretrained(base, adapter_id)
- Created
- Sep 5, 2026
- Updated
- Sep 5, 2026