fine-tuned-DatasetQAS-IDK-MRC-with-indobert-large-p2-without-ITTL-without-freeze-LR-1e-05

This model is a fine-tuned version of indobenchmark/indobert-large-p2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.2405
• Exact Match: 51.9634
• F1: 58.9740
• Precision: 59.9859
• Recall: 63.1982

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 16

Training results

Training Loss	Epoch	Step	Validation Loss	Exact Match	F1	Precision	Recall
3.4722	0.49	73	2.4443	9.6859	19.2463	17.0064	36.4763
2.4592	0.99	146	1.8046	26.4398	35.0647	34.6691	45.8354
1.6685	1.49	219	1.3839	42.0157	49.2034	49.8767	57.6434
1.4304	1.98	292	1.3337	42.4084	50.0207	51.1139	56.5044
1.074	2.48	365	1.2313	49.6073	56.5977	57.6801	61.6748
1.0704	2.97	438	1.1639	52.0942	58.7433	59.7956	64.2988
0.8772	3.47	511	1.1926	53.6649	61.0220	61.4728	66.7934
0.8887	3.97	584	1.2182	51.0471	58.0581	59.1383	62.9618
0.7141	4.47	657	1.1726	54.3194	61.1547	62.0500	66.0011
0.7238	4.96	730	1.2156	54.0576	60.4732	61.7775	64.5058
0.5929	5.46	803	1.3549	52.7487	59.0996	60.3930	62.9242
0.6201	5.95	876	1.2405	51.9634	58.9740	59.9859	63.1982

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1+cu117
• Datasets 2.2.0
• Tokenizers 0.13.2