arabic-hs-2class-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: 0.6540
• Accuracy: 0.8442
• Macro F1: 0.8331

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
0.6218	0.1083	100	0.5720	0.6917	0.5951
0.5615	0.2167	200	0.5156	0.7297	0.6866
0.5132	0.3250	300	0.4785	0.7554	0.7400
0.4878	0.4334	400	0.4444	0.7879	0.7657
0.4623	0.5417	500	0.4162	0.8049	0.7858
0.4643	0.6501	600	0.4030	0.8042	0.7923
0.4455	0.7584	700	0.4118	0.8178	0.7969
0.4406	0.8667	800	0.3787	0.8320	0.8176
0.4145	0.9751	900	0.3786	0.8252	0.8136
0.4014	1.0834	1000	0.3790	0.8340	0.8174
0.3784	1.1918	1100	0.3703	0.8401	0.8285
0.3858	1.3001	1200	0.3651	0.8388	0.8290
0.3721	1.4085	1300	0.3605	0.8435	0.8323
0.3836	1.5168	1400	0.3604	0.8394	0.8312
0.3645	1.6251	1500	0.3658	0.8489	0.8343
0.375	1.7335	1600	0.3790	0.8347	0.8144
0.3909	1.8418	1700	0.3526	0.8469	0.8374
0.3566	1.9502	1800	0.3701	0.8482	0.8300
0.3243	2.0585	1900	0.3598	0.8388	0.8310
0.3699	2.1668	2000	0.3753	0.8340	0.8108
0.32	2.2752	2100	0.3541	0.8523	0.8408
0.2987	2.3835	2200	0.3588	0.8469	0.8370
0.3037	2.4919	2300	0.3687	0.8523	0.8398
0.3156	2.6002	2400	0.3522	0.8469	0.8378
0.3352	2.7086	2500	0.3440	0.8503	0.8415
0.3225	2.8169	2600	0.3490	0.8489	0.8394
0.3	2.9252	2700	0.3638	0.8476	0.8328
0.2795	3.0336	2800	0.3904	0.8482	0.8334
0.2787	3.1419	2900	0.4356	0.8415	0.8194
0.2733	3.2503	3000	0.3792	0.8530	0.8435
0.2765	3.3586	3100	0.3700	0.8496	0.8412
0.2746	3.4670	3200	0.3816	0.8476	0.8343
0.2823	3.5753	3300	0.3704	0.8503	0.8374
0.2726	3.6836	3400	0.3795	0.8516	0.8406
0.2661	3.7920	3500	0.4218	0.8469	0.8283
0.265	3.9003	3600	0.3852	0.8523	0.8425
0.2686	4.0087	3700	0.3782	0.8516	0.8428
0.2173	4.1170	3800	0.4008	0.8496	0.8401
0.2404	4.2254	3900	0.4072	0.8577	0.8453
0.2635	4.3337	4000	0.3811	0.8564	0.8464
0.1983	4.4420	4100	0.4038	0.8598	0.8506
0.2512	4.5504	4200	0.4228	0.8482	0.8343
0.2375	4.6587	4300	0.4070	0.8476	0.8359
0.2405	4.7671	4400	0.4433	0.8415	0.8242
0.2319	4.8754	4500	0.4176	0.8469	0.8370
0.2249	4.9837	4600	0.4342	0.8462	0.8361
0.1937	5.0921	4700	0.4543	0.8476	0.8369
0.2207	5.2004	4800	0.4553	0.8489	0.8357
0.1963	5.3088	4900	0.4779	0.8489	0.8331
0.178	5.4171	5000	0.4590	0.8455	0.8344
0.175	5.5255	5100	0.4663	0.8523	0.8421
0.1999	5.6338	5200	0.4733	0.8523	0.8413
0.2018	5.7421	5300	0.4649	0.8523	0.8411
0.208	5.8505	5400	0.4617	0.8482	0.8368
0.2096	5.9588	5500	0.4561	0.8482	0.8379
0.1972	6.0672	5600	0.5073	0.8476	0.8331
0.1815	6.1755	5700	0.4908	0.8469	0.8358
0.1625	6.2839	5800	0.5171	0.8435	0.8298
0.1869	6.3922	5900	0.5083	0.8462	0.8346
0.1672	6.5005	6000	0.4982	0.8509	0.8405
0.1757	6.6089	6100	0.5455	0.8462	0.8307
0.1505	6.7172	6200	0.5220	0.8428	0.8313
0.1455	6.8256	6300	0.5378	0.8428	0.8318
0.2074	6.9339	6400	0.5219	0.8442	0.8336
0.1551	7.0423	6500	0.5336	0.8394	0.8287
0.1501	7.1506	6600	0.5581	0.8435	0.8323
0.167	7.2589	6700	0.5594	0.8408	0.8290
0.1559	7.3673	6800	0.5704	0.8449	0.8313
0.1676	7.4756	6900	0.5758	0.8455	0.8319
0.134	7.5840	7000	0.5662	0.8482	0.8371
0.1773	7.6923	7100	0.5677	0.8435	0.8323
0.1328	7.8007	7200	0.5843	0.8482	0.8363
0.1241	7.9090	7300	0.5901	0.8435	0.8332
0.1657	8.0173	7400	0.6129	0.8374	0.8281
0.1348	8.1257	7500	0.6221	0.8442	0.8340
0.1414	8.2340	7600	0.6272	0.8421	0.8319
0.1166	8.3424	7700	0.6430	0.8428	0.8324
0.1648	8.4507	7800	0.6359	0.8489	0.8376
0.1359	8.5590	7900	0.6338	0.8476	0.8367
0.16	8.6674	8000	0.6433	0.8421	0.8293
0.1256	8.7757	8100	0.6438	0.8455	0.8339
0.1673	8.8841	8200	0.6452	0.8428	0.8304
0.151	8.9924	8300	0.6377	0.8394	0.8281
0.1381	9.1008	8400	0.6409	0.8442	0.8318
0.1199	9.2091	8500	0.6336	0.8408	0.8303
0.1052	9.3174	8600	0.6434	0.8401	0.8286
0.14	9.4258	8700	0.6460	0.8421	0.8310
0.1357	9.5341	8800	0.6503	0.8428	0.8310
0.1149	9.6425	8900	0.6580	0.8428	0.8301
0.1486	9.7508	9000	0.6552	0.8442	0.8322
0.1315	9.8592	9100	0.6518	0.8442	0.8335
0.1125	9.9675	9200	0.6540	0.8442	0.8331

Framework versions

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