metadata

license: mit
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - f1
  - precision
  - recall
model-index:
  - name: roberta-news-classifier
    results: []

roberta-news-classifier

This model is a fine-tuned version of burakaytan/roberta-base-turkish-uncased on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.2394
Accuracy: 0.9388
F1: 0.9388
Precision: 0.9388
Recall: 0.9388

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: 1e-05
train_batch_size: 64
eval_batch_size: 150
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 500
num_epochs: 12

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.2929	1.0	62	0.2893	0.9316	0.9316	0.9316	0.9316
0.2775	2.0	124	0.2700	0.9337	0.9337	0.9337	0.9337
0.2554	3.0	186	0.2576	0.9286	0.9286	0.9286	0.9286
0.2198	4.0	248	0.2409	0.9286	0.9286	0.9286	0.9286
0.197	5.0	310	0.2324	0.9306	0.9306	0.9306	0.9306
0.1611	6.0	372	0.2254	0.9367	0.9367	0.9367	0.9367
0.1302	7.0	434	0.2204	0.9378	0.9378	0.9378	0.9378
0.1058	8.0	496	0.2238	0.9337	0.9337	0.9337	0.9337
0.0976	9.0	558	0.2295	0.9378	0.9378	0.9378	0.9378
0.0795	10.0	620	0.2299	0.9378	0.9378	0.9378	0.9378
0.0641	11.0	682	0.2394	0.9388	0.9388	0.9388	0.9388
0.0544	12.0	744	0.2392	0.9367	0.9367	0.9367	0.9367

Framework versions

Transformers 4.24.0
Pytorch 1.12.1+cu113
Datasets 2.7.1
Tokenizers 0.13.2