arabic-hs-group-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: 0.3012
• Accuracy: 0.9444
• Macro F1: 0.9147

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
2.07	0.1029	100	1.9058	0.4144	0.1548
1.7855	0.2058	200	1.6273	0.4907	0.2162
1.539	0.3086	300	1.4125	0.5683	0.2940
1.2834	0.4115	400	1.1540	0.6586	0.3591
1.1157	0.5144	500	0.9695	0.7095	0.4362
0.9701	0.6173	600	0.8559	0.7488	0.5098
0.848	0.7202	700	0.8231	0.7714	0.5732
0.7585	0.8230	800	0.6777	0.8171	0.6513
0.698	0.9259	900	0.6211	0.8281	0.6726
0.6143	1.0288	1000	0.5631	0.8466	0.7037
0.5656	1.1317	1100	0.5575	0.8519	0.7296
0.4638	1.2346	1200	0.4738	0.8738	0.7532
0.4572	1.3374	1300	0.4279	0.8877	0.7689
0.3826	1.4403	1400	0.4296	0.8814	0.7680
0.3424	1.5432	1500	0.3643	0.9045	0.7900
0.3636	1.6461	1600	0.3541	0.9057	0.7895
0.3061	1.7490	1700	0.3264	0.9120	0.7982
0.3145	1.8519	1800	0.3232	0.9126	0.7974
0.3074	1.9547	1900	0.3124	0.9167	0.8140
0.3145	2.0576	2000	0.3228	0.9144	0.8238
0.2337	2.1605	2100	0.3152	0.9167	0.8236
0.2283	2.2634	2200	0.3145	0.9184	0.8286
0.2265	2.3663	2300	0.3230	0.9207	0.8212
0.2428	2.4691	2400	0.3038	0.9248	0.8434
0.2301	2.5720	2500	0.2946	0.9225	0.8523
0.2471	2.6749	2600	0.2937	0.9201	0.8516
0.1854	2.7778	2700	0.2864	0.9288	0.8594
0.2154	2.8807	2800	0.2896	0.9259	0.8526
0.2196	2.9835	2900	0.2891	0.9271	0.8727
0.1772	3.0864	3000	0.2884	0.9271	0.8537
0.1876	3.1893	3100	0.2827	0.9282	0.8730
0.1898	3.2922	3200	0.2698	0.9300	0.8584
0.1592	3.3951	3300	0.2923	0.9334	0.8752
0.1423	3.4979	3400	0.2738	0.9329	0.8862
0.1818	3.6008	3500	0.2811	0.9294	0.8825
0.1357	3.7037	3600	0.2758	0.9381	0.8940
0.1667	3.8066	3700	0.3017	0.9317	0.8941
0.1609	3.9095	3800	0.2881	0.9340	0.8914
0.141	4.0123	3900	0.3014	0.9329	0.9001
0.0912	4.1152	4000	0.2830	0.9346	0.9008
0.1152	4.2181	4100	0.3018	0.9329	0.8922
0.1576	4.3210	4200	0.2790	0.9375	0.9090
0.1255	4.4239	4300	0.2893	0.9329	0.8968
0.1186	4.5267	4400	0.3274	0.9265	0.8982
0.1362	4.6296	4500	0.2865	0.9363	0.9025
0.1171	4.7325	4600	0.2941	0.9358	0.9042
0.1494	4.8354	4700	0.2841	0.9352	0.9006
0.1065	4.9383	4800	0.3263	0.9300	0.9010
0.104	5.0412	4900	0.2949	0.9398	0.9146
0.0767	5.1440	5000	0.3041	0.9375	0.9025
0.0924	5.2469	5100	0.2908	0.9381	0.9125
0.1106	5.3498	5200	0.3003	0.9387	0.9006
0.1099	5.4527	5300	0.2844	0.9410	0.9084
0.0985	5.5556	5400	0.2936	0.9363	0.9070
0.087	5.6584	5500	0.2828	0.9404	0.9089
0.0997	5.7613	5600	0.2881	0.9387	0.9119
0.0982	5.8642	5700	0.2950	0.9363	0.9087
0.1242	5.9671	5800	0.3027	0.9398	0.9074
0.0947	6.0700	5900	0.2926	0.9398	0.9115
0.0845	6.1728	6000	0.2823	0.9427	0.9101
0.0881	6.2757	6100	0.3071	0.9363	0.9144
0.0806	6.3786	6200	0.2972	0.9392	0.9101
0.0676	6.4815	6300	0.2839	0.9381	0.8999
0.1106	6.5844	6400	0.3066	0.9358	0.9020
0.1031	6.6872	6500	0.2971	0.9387	0.9034
0.0691	6.7901	6600	0.2932	0.9416	0.9130
0.0779	6.8930	6700	0.3020	0.9410	0.9121
0.0676	6.9959	6800	0.2931	0.9416	0.9144
0.0721	7.0988	6900	0.2877	0.9398	0.9114
0.0552	7.2016	7000	0.2978	0.9450	0.9241
0.0502	7.3045	7100	0.2899	0.9416	0.9122
0.0572	7.4074	7200	0.2941	0.9421	0.9131
0.079	7.5103	7300	0.2913	0.9427	0.9161
0.0676	7.6132	7400	0.2965	0.9410	0.9135
0.0673	7.7160	7500	0.3055	0.9404	0.9076
0.0709	7.8189	7600	0.3005	0.9421	0.9122
0.0709	7.9218	7700	0.3023	0.9416	0.9160
0.0679	8.0247	7800	0.2973	0.9410	0.9108
0.0542	8.1276	7900	0.2913	0.9427	0.9174
0.0605	8.2305	8000	0.3070	0.9421	0.9154
0.0578	8.3333	8100	0.3053	0.9410	0.9140
0.0457	8.4362	8200	0.2946	0.9450	0.9214
0.0631	8.5391	8300	0.3010	0.9433	0.9094
0.0563	8.6420	8400	0.2957	0.9433	0.9177
0.0624	8.7449	8500	0.2965	0.9468	0.9209
0.0557	8.8477	8600	0.3052	0.9416	0.9135
0.058	8.9506	8700	0.3091	0.9416	0.9129
0.0319	9.0535	8800	0.3078	0.9439	0.9156
0.0454	9.1564	8900	0.3069	0.9433	0.9109
0.0583	9.2593	9000	0.2996	0.9450	0.9156
0.0668	9.3621	9100	0.3026	0.9427	0.9132
0.0615	9.4650	9200	0.3065	0.9421	0.9115
0.0506	9.5679	9300	0.3037	0.9439	0.9150
0.0514	9.6708	9400	0.3012	0.9450	0.9164
0.0447	9.7737	9500	0.3004	0.9450	0.9156
0.0532	9.8765	9600	0.3009	0.9444	0.9147
0.0568	9.9794	9700	0.3012	0.9444	0.9147

Framework versions

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