README.md

metadata

library_name: transformers
license: mit
base_model: roberta-base
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - f1
  - precision
  - recall
model-index:
  - name: checkpoints
    results: []

Qna classifier

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

• Loss: 0.4574
• Accuracy: 0.8526
• F1: 0.8402
• Precision: 0.8566
• Recall: 0.8526

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
No log	0.1485	500	3.3156	0.2530	0.1877	0.1772	0.2530
No log	0.2971	1000	2.5424	0.3081	0.2274	0.2078	0.3081
No log	0.4456	1500	2.1390	0.3414	0.2649	0.2471	0.3414
No log	0.5942	2000	1.8717	0.4001	0.3334	0.3560	0.4001
No log	0.7427	2500	1.6311	0.4522	0.3897	0.3845	0.4522
No log	0.8913	3000	1.4911	0.5069	0.4460	0.4512	0.5069
2.3757	1.0398	3500	1.3221	0.5510	0.4840	0.4818	0.5510
2.3757	1.1884	4000	1.1827	0.6173	0.5630	0.5779	0.6173
2.3757	1.3369	4500	1.0504	0.6518	0.6041	0.6272	0.6518
2.3757	1.4854	5000	0.9556	0.6752	0.6272	0.6453	0.6752
2.3757	1.6340	5500	0.8580	0.7113	0.6642	0.6795	0.7113
2.3757	1.7825	6000	0.7818	0.7364	0.6984	0.7270	0.7364
2.3757	1.9311	6500	0.7009	0.7669	0.7324	0.7616	0.7669
1.0183	2.0796	7000	0.6271	0.7957	0.7685	0.7883	0.7957
1.0183	2.2282	7500	0.5864	0.8051	0.7812	0.8061	0.8051
1.0183	2.3767	8000	0.5439	0.8237	0.8008	0.8104	0.8237
1.0183	2.5253	8500	0.5116	0.8297	0.8118	0.8268	0.8297
1.0183	2.6738	9000	0.4916	0.8394	0.8237	0.8450	0.8394
1.0183	2.8223	9500	0.4696	0.8486	0.8346	0.8514	0.8486
1.0183	2.9709	10000	0.4574	0.8526	0.8402	0.8566	0.8526

Framework versions

• Transformers 4.53.1
• Pytorch 2.6.0+cu124
• Datasets 2.14.4
• Tokenizers 0.21.2