Türkçe Multi-label Intent Classification RoBERTa
Depremzedelerin ihtiyaçlarını karşılamak için etiketlenmiş eğitilmiş multi-label RoBERTa modeli. Aşağıda değerlendirme sonuçları var.
Evaluation
'eval_loss': 0.18568251545368838,
'eval_runtime': 2.7693,
'eval_samples_per_second': 254.935,
'eval_steps_per_second': 8.305,
'epoch': 3.0
Classification Report
precision recall f1-score support
Alakasiz 0.95 0.87 0.91 781
Barinma 0.86 0.52 0.65 234
Elektronik 0.00 0.00 0.00 171
Giysi 0.89 0.25 0.39 122
Kurtarma 0.86 0.78 0.82 472
Lojistik 0.00 0.00 0.00 123
Saglik 0.78 0.05 0.09 148
Su 0.92 0.11 0.20 96
Yagma 0.00 0.00 0.00 19
Yemek 0.94 0.42 0.58 158
micro avg 0.91 0.55 0.69 2324
macro avg 0.62 0.30 0.36 2324
weighted avg 0.78 0.55 0.61 2324
samples avg 0.69 0.63 0.65 2324
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Spaces using deprem-ml/deprem-roberta-intent 2
Evaluation results
- recall on deprem_private_dataset_v1_2self-reported0.620
- f1 on deprem_private_dataset_v1_2self-reported0.360