johannawawi/1_model-for-sosmed-analysis-v3

results_final

This model is a fine-tuned version of w11wo/indonesian-roberta-base-sentiment-classifier on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4292
• Accuracy: 0.8881
• F1 Macro: 0.8880
• F1 Weighted: 0.8880
• Precision Macro: 0.8896
• Recall Macro: 0.8881
• Precision Weighted: 0.8895
• Recall Weighted: 0.8881

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: 3e-06
• 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: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro	F1 Weighted	Precision Macro	Recall Macro	Precision Weighted	Recall Weighted
0.2316	0.3436	100	0.4071	0.8864	0.8862	0.8862	0.8869	0.8864	0.8870	0.8864
0.2298	0.6873	200	0.4028	0.8864	0.8863	0.8863	0.8867	0.8864	0.8867	0.8864
0.1311	1.0309	300	0.4144	0.8847	0.8846	0.8846	0.8852	0.8847	0.8852	0.8847
0.159	1.3746	400	0.4292	0.8881	0.8880	0.8880	0.8896	0.8881	0.8895	0.8881
0.1957	1.7182	500	0.4283	0.8830	0.8828	0.8829	0.8835	0.8829	0.8836	0.8830
0.228	2.0619	600	0.4153	0.8778	0.8778	0.8779	0.8783	0.8778	0.8783	0.8778
0.2248	2.4055	700	0.4242	0.8830	0.8828	0.8828	0.8833	0.8829	0.8833	0.8830
0.1733	2.7491	800	0.4239	0.8795	0.8795	0.8795	0.8803	0.8795	0.8803	0.8795
0.2314	3.0928	900	0.4166	0.8812	0.8811	0.8811	0.8813	0.8812	0.8813	0.8812
0.1691	3.4364	1000	0.4472	0.8744	0.8741	0.8741	0.8757	0.8743	0.8757	0.8744
0.2671	3.7801	1100	0.4273	0.8830	0.8828	0.8829	0.8832	0.8829	0.8832	0.8830
0.2643	4.1237	1200	0.4317	0.8812	0.8811	0.8811	0.8815	0.8812	0.8815	0.8812
0.202	4.4674	1300	0.4440	0.8847	0.8846	0.8846	0.8852	0.8846	0.8852	0.8847
0.2538	4.8110	1400	0.4397	0.8812	0.8812	0.8812	0.8816	0.8812	0.8816	0.8812
0.2662	5.1546	1500	0.4364	0.8847	0.8846	0.8846	0.8852	0.8846	0.8852	0.8847
0.2655	5.4983	1600	0.4298	0.8812	0.8812	0.8812	0.8816	0.8812	0.8816	0.8812
0.1933	5.8419	1700	0.4422	0.8847	0.8845	0.8845	0.8854	0.8846	0.8854	0.8847
0.2289	6.1856	1800	0.4282	0.8778	0.8776	0.8777	0.8778	0.8778	0.8778	0.8778
0.2298	6.5292	1900	0.4313	0.8795	0.8794	0.8794	0.8798	0.8795	0.8798	0.8795
0.2008	6.8729	2000	0.4344	0.8812	0.8811	0.8812	0.8816	0.8812	0.8817	0.8812
0.2107	7.2165	2100	0.4354	0.8830	0.8829	0.8829	0.8833	0.8829	0.8834	0.8830
0.2505	7.5601	2200	0.4353	0.8830	0.8829	0.8829	0.8833	0.8829	0.8834	0.8830
0.2134	7.9038	2300	0.4361	0.8830	0.8829	0.8829	0.8833	0.8829	0.8834	0.8830
0.2613	8.2474	2400	0.4344	0.8830	0.8828	0.8829	0.8832	0.8829	0.8832	0.8830
0.2128	8.5911	2500	0.4350	0.8830	0.8828	0.8829	0.8832	0.8829	0.8832	0.8830
0.212	8.9347	2600	0.4356	0.8830	0.8828	0.8829	0.8832	0.8829	0.8832	0.8830
0.2379	9.2784	2700	0.4359	0.8830	0.8828	0.8829	0.8832	0.8829	0.8832	0.8830
0.1471	9.6220	2800	0.4358	0.8830	0.8828	0.8829	0.8832	0.8829	0.8832	0.8830
0.1781	9.9656	2900	0.4359	0.8830	0.8828	0.8829	0.8832	0.8829	0.8832	0.8830

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 2.14.4
• Tokenizers 0.21.1