arabert_cross_relevance_task7_fold6

This model is a fine-tuned version of aubmindlab/bert-base-arabertv02 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4397
• Qwk: 0.1488
• Mse: 0.4393

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: 2e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Qwk	Mse
No log	0.125	2	0.7246	0.0092	0.7226
No log	0.25	4	0.3923	0.1169	0.3915
No log	0.375	6	0.3543	0.0420	0.3537
No log	0.5	8	0.3794	0.1151	0.3788
No log	0.625	10	0.3085	0.0822	0.3084
No log	0.75	12	0.2791	0.0429	0.2796
No log	0.875	14	0.2793	0.0732	0.2799
No log	1.0	16	0.2798	0.0822	0.2804
No log	1.125	18	0.2748	0.1450	0.2754
No log	1.25	20	0.2618	0.1769	0.2623
No log	1.375	22	0.2610	0.2119	0.2614
No log	1.5	24	0.2714	0.2099	0.2716
No log	1.625	26	0.2877	0.2282	0.2877
No log	1.75	28	0.3085	0.2170	0.3085
No log	1.875	30	0.2818	0.2521	0.2819
No log	2.0	32	0.2704	0.2285	0.2706
No log	2.125	34	0.2630	0.2177	0.2635
No log	2.25	36	0.2577	0.1930	0.2582
No log	2.375	38	0.2575	0.2186	0.2579
No log	2.5	40	0.2613	0.2282	0.2617
No log	2.625	42	0.2766	0.2044	0.2769
No log	2.75	44	0.2809	0.2091	0.2811
No log	2.875	46	0.2689	0.2194	0.2691
No log	3.0	48	0.2629	0.2198	0.2632
No log	3.125	50	0.2671	0.2133	0.2672
No log	3.25	52	0.2896	0.2170	0.2895
No log	3.375	54	0.3099	0.2118	0.3097
No log	3.5	56	0.2867	0.2181	0.2867
No log	3.625	58	0.2681	0.2108	0.2683
No log	3.75	60	0.2670	0.2042	0.2674
No log	3.875	62	0.2777	0.2220	0.2780
No log	4.0	64	0.2823	0.2220	0.2826
No log	4.125	66	0.2925	0.2150	0.2927
No log	4.25	68	0.3165	0.2285	0.3164
No log	4.375	70	0.3340	0.2156	0.3337
No log	4.5	72	0.3690	0.1876	0.3685
No log	4.625	74	0.3812	0.1800	0.3806
No log	4.75	76	0.3599	0.1954	0.3594
No log	4.875	78	0.3297	0.2097	0.3294
No log	5.0	80	0.3082	0.2063	0.3081
No log	5.125	82	0.2990	0.2193	0.2991
No log	5.25	84	0.3004	0.2193	0.3005
No log	5.375	86	0.3010	0.2140	0.3011
No log	5.5	88	0.3297	0.1814	0.3298
No log	5.625	90	0.3943	0.1785	0.3943
No log	5.75	92	0.4229	0.1592	0.4228
No log	5.875	94	0.3892	0.1719	0.3892
No log	6.0	96	0.3418	0.1836	0.3418
No log	6.125	98	0.3361	0.1932	0.3360
No log	6.25	100	0.3641	0.1930	0.3638
No log	6.375	102	0.3934	0.1844	0.3930
No log	6.5	104	0.3956	0.1915	0.3953
No log	6.625	106	0.3700	0.2006	0.3700
No log	6.75	108	0.3683	0.1800	0.3684
No log	6.875	110	0.3757	0.1785	0.3758
No log	7.0	112	0.4066	0.1688	0.4066
No log	7.125	114	0.4152	0.1695	0.4151
No log	7.25	116	0.3917	0.1790	0.3917
No log	7.375	118	0.3540	0.1833	0.3542
No log	7.5	120	0.3449	0.1998	0.3451
No log	7.625	122	0.3587	0.1955	0.3588
No log	7.75	124	0.3969	0.1807	0.3968
No log	7.875	126	0.4580	0.1405	0.4577
No log	8.0	128	0.5018	0.1433	0.5013
No log	8.125	130	0.5002	0.1433	0.4997
No log	8.25	132	0.4689	0.1442	0.4684
No log	8.375	134	0.4386	0.1488	0.4382
No log	8.5	136	0.4174	0.1550	0.4171
No log	8.625	138	0.4152	0.1613	0.4149
No log	8.75	140	0.4293	0.1550	0.4290
No log	8.875	142	0.4512	0.1427	0.4508
No log	9.0	144	0.4641	0.1405	0.4637
No log	9.125	146	0.4566	0.1427	0.4562
No log	9.25	148	0.4432	0.1427	0.4429
No log	9.375	150	0.4330	0.1550	0.4327
No log	9.5	152	0.4301	0.1613	0.4298
No log	9.625	154	0.4343	0.1550	0.4340
No log	9.75	156	0.4358	0.1550	0.4354
No log	9.875	158	0.4385	0.1488	0.4381
No log	10.0	160	0.4397	0.1488	0.4393

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0
• Datasets 2.21.0
• Tokenizers 0.19.1