arabic-hs-4class-prediction

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

• Loss: 0.7358
• Accuracy: 0.8029
• Macro F1: 0.6756

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-06
• train_batch_size: 16
• eval_batch_size: 20
• 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: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Macro F1
0.9671	0.1147	100	0.8356	0.7075	0.3086
0.8352	0.2294	200	0.7748	0.7312	0.3898
0.7784	0.3440	300	0.7342	0.7341	0.3953
0.6864	0.4587	400	0.6894	0.7491	0.4471
0.7008	0.5734	500	0.6578	0.7763	0.5418
0.6343	0.6881	600	0.6414	0.7692	0.5089
0.6256	0.8028	700	0.6297	0.7699	0.5101
0.6397	0.9174	800	0.6173	0.7857	0.5482
0.6386	1.0321	900	0.6079	0.7821	0.5324
0.5845	1.1468	1000	0.6030	0.7799	0.5436
0.5638	1.2615	1100	0.5884	0.7735	0.5558
0.5811	1.3761	1200	0.5954	0.7885	0.5616
0.5892	1.4908	1300	0.5859	0.7900	0.6102
0.5539	1.6055	1400	0.5773	0.7871	0.6078
0.5866	1.7202	1500	0.5779	0.7935	0.6306
0.5884	1.8349	1600	0.5746	0.7885	0.6056
0.5502	1.9495	1700	0.5752	0.7935	0.6032
0.5369	2.0642	1800	0.5732	0.7928	0.6303
0.4772	2.1789	1900	0.5766	0.7928	0.6170
0.5344	2.2936	2000	0.5679	0.7978	0.6329
0.4929	2.4083	2100	0.5776	0.7821	0.6099
0.4743	2.5229	2200	0.6351	0.7978	0.6143
0.5125	2.6376	2300	0.5809	0.8014	0.6551
0.4917	2.7523	2400	0.5674	0.8007	0.6275
0.4894	2.8670	2500	0.5637	0.7907	0.6383
0.4739	2.9817	2600	0.5618	0.7971	0.6560
0.4364	3.0963	2700	0.5690	0.7964	0.6464
0.4021	3.2110	2800	0.5883	0.8043	0.6484
0.4382	3.3257	2900	0.6049	0.8086	0.6460
0.4441	3.4404	3000	0.5804	0.7950	0.6571
0.4514	3.5550	3100	0.6004	0.7842	0.6288
0.4783	3.6697	3200	0.5746	0.7921	0.6420
0.4358	3.7844	3300	0.5769	0.7957	0.6580
0.405	3.8991	3400	0.5888	0.8050	0.6580
0.4349	4.0138	3500	0.5718	0.8072	0.6692
0.3575	4.1284	3600	0.6027	0.7907	0.6561
0.3965	4.2431	3700	0.6006	0.7971	0.6677
0.396	4.3578	3800	0.6009	0.7928	0.6564
0.3564	4.4725	3900	0.6015	0.8043	0.6598
0.3921	4.5872	4000	0.6052	0.7978	0.6649
0.4333	4.7018	4100	0.6017	0.8029	0.6585
0.3763	4.8165	4200	0.6016	0.8007	0.6668
0.3518	4.9312	4300	0.6034	0.7950	0.6567
0.3347	5.0459	4400	0.6364	0.7921	0.6690
0.337	5.1606	4500	0.6507	0.8093	0.6680
0.3537	5.2752	4600	0.6392	0.8	0.6683
0.3433	5.3899	4700	0.6250	0.8	0.6714
0.3465	5.5046	4800	0.6334	0.7978	0.6742
0.3127	5.6193	4900	0.6433	0.7986	0.6716
0.3416	5.7339	5000	0.6328	0.7943	0.6629
0.3339	5.8486	5100	0.6271	0.8014	0.6708
0.3382	5.9633	5200	0.6418	0.7964	0.6684
0.3226	6.0780	5300	0.6600	0.7935	0.6721
0.3346	6.1927	5400	0.6494	0.7921	0.6724
0.3074	6.3073	5500	0.6533	0.7964	0.6795
0.2975	6.4220	5600	0.6606	0.7928	0.6693
0.3047	6.5367	5700	0.6683	0.8	0.6709
0.2818	6.6514	5800	0.6797	0.8022	0.6742
0.3164	6.7661	5900	0.6804	0.7950	0.6664
0.2959	6.8807	6000	0.6814	0.7957	0.6596
0.2941	6.9954	6100	0.6810	0.7935	0.6711
0.2954	7.1101	6200	0.6790	0.7892	0.6578
0.2615	7.2248	6300	0.6998	0.7993	0.6605
0.2395	7.3394	6400	0.7026	0.7957	0.6661
0.3184	7.4541	6500	0.7183	0.7785	0.6564
0.3012	7.5688	6600	0.6923	0.7921	0.6659
0.2446	7.6835	6700	0.6981	0.7986	0.6660
0.2895	7.7982	6800	0.6853	0.8014	0.6743
0.2854	7.9128	6900	0.6916	0.7978	0.6704
0.2599	8.0275	7000	0.6993	0.7957	0.6713
0.2405	8.1422	7100	0.7094	0.7935	0.6695
0.2445	8.2569	7200	0.7101	0.7950	0.6706
0.2537	8.3716	7300	0.7169	0.8036	0.6707
0.2573	8.4862	7400	0.7095	0.7971	0.6744
0.2316	8.6009	7500	0.7215	0.8007	0.6729
0.2726	8.7156	7600	0.7232	0.7971	0.6743
0.2371	8.8303	7700	0.7227	0.7964	0.6704
0.2554	8.9450	7800	0.7217	0.7986	0.6714
0.2284	9.0596	7900	0.7243	0.8036	0.6776
0.2442	9.1743	8000	0.7305	0.8022	0.6759
0.2369	9.2890	8100	0.7322	0.8022	0.6767
0.2769	9.4037	8200	0.7324	0.8057	0.6822
0.2417	9.5183	8300	0.7314	0.8007	0.6745
0.2529	9.6330	8400	0.7333	0.7971	0.6737
0.2441	9.7477	8500	0.7341	0.7964	0.6719
0.2272	9.8624	8600	0.7365	0.8014	0.6722
0.2208	9.9771	8700	0.7358	0.8029	0.6756

Framework versions

• Transformers 4.49.0
• Pytorch 2.5.1+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0