turkish-zeroshot

This model is a fine-tuned version of dbmdz/bert-base-turkish-cased onfacebook/xnli tr dataset. It achieves the following results on the evaluation set:

• Loss: 0.5637
• Accuracy: 0.7731
• F1: 0.7740
• Precision: 0.7804
• Recall: 0.7731

Usage

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("zero-shot-classification", model="kaixkhazaki/turkish-zeroshot")

#Enter your text and possible candidates of classification
sequence = "Bu laptopun pil ömrü ne kadar dayanıyor?"
candidate_labels = ["ürün özellikleri", "soru", "bilgi talebi", "laptop", "teknik destek"]

pipe(
    sequence,
    candidate_labels,
)

>>
{'sequence': 'Bu laptopun pil ömrü ne kadar dayanıyor?',
 'labels': ['ürün özellikleri',
  'soru',
  'bilgi talebi',
  'laptop',
  'teknik destek'],
 'scores': [0.296932578086853,
  0.2693993151187897,
  0.20735479891300201,
  0.12200483679771423,
  0.10430848598480225]}

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 64
• eval_batch_size: 32
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 500
• num_epochs: 5

Training results

Framework versions

• Transformers 4.48.0.dev0
• Pytorch 2.4.1+cu121
• Datasets 3.1.0
• Tokenizers 0.21.0