arabert_cross_relevance_task2_fold4

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: 0.4006
• Qwk: 0.2250
• Mse: 0.4006

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	1.1063	0.0234	1.1063
No log	0.25	4	0.4473	0.1039	0.4473
No log	0.375	6	0.4088	0.2073	0.4088
No log	0.5	8	0.4319	0.2351	0.4319
No log	0.625	10	0.2875	0.2401	0.2875
No log	0.75	12	0.2836	0.2194	0.2836
No log	0.875	14	0.3021	0.2575	0.3021
No log	1.0	16	0.2942	0.2831	0.2942
No log	1.125	18	0.3155	0.2508	0.3155
No log	1.25	20	0.2964	0.2787	0.2964
No log	1.375	22	0.2774	0.2820	0.2774
No log	1.5	24	0.2398	0.3029	0.2398
No log	1.625	26	0.2524	0.2501	0.2524
No log	1.75	28	0.2680	0.2390	0.2680
No log	1.875	30	0.2581	0.2579	0.2581
No log	2.0	32	0.2540	0.2787	0.2540
No log	2.125	34	0.2836	0.3308	0.2836
No log	2.25	36	0.3870	0.2114	0.3870
No log	2.375	38	0.4205	0.2269	0.4205
No log	2.5	40	0.3210	0.2236	0.3210
No log	2.625	42	0.2478	0.4116	0.2478
No log	2.75	44	0.2382	0.3767	0.2382
No log	2.875	46	0.2494	0.2579	0.2494
No log	3.0	48	0.3000	0.2553	0.3000
No log	3.125	50	0.3533	0.2028	0.3533
No log	3.25	52	0.3681	0.2209	0.3681
No log	3.375	54	0.3235	0.2270	0.3235
No log	3.5	56	0.2844	0.2648	0.2844
No log	3.625	58	0.2941	0.2428	0.2941
No log	3.75	60	0.3737	0.2222	0.3737
No log	3.875	62	0.4306	0.2146	0.4306
No log	4.0	64	0.4279	0.2146	0.4279
No log	4.125	66	0.3959	0.2155	0.3959
No log	4.25	68	0.3136	0.2441	0.3136
No log	4.375	70	0.2972	0.2593	0.2972
No log	4.5	72	0.2889	0.2677	0.2889
No log	4.625	74	0.3100	0.2489	0.3100
No log	4.75	76	0.3614	0.2425	0.3614
No log	4.875	78	0.3949	0.2337	0.3949
No log	5.0	80	0.4746	0.1752	0.4746
No log	5.125	82	0.4462	0.1969	0.4462
No log	5.25	84	0.3424	0.2572	0.3424
No log	5.375	86	0.3171	0.2798	0.3171
No log	5.5	88	0.3687	0.2452	0.3687
No log	5.625	90	0.4496	0.1969	0.4496
No log	5.75	92	0.4134	0.2126	0.4134
No log	5.875	94	0.3763	0.2172	0.3763
No log	6.0	96	0.3924	0.2106	0.3924
No log	6.125	98	0.3918	0.2106	0.3918
No log	6.25	100	0.4398	0.2029	0.4398
No log	6.375	102	0.5008	0.1879	0.5008
No log	6.5	104	0.4898	0.1781	0.4898
No log	6.625	106	0.4198	0.2218	0.4198
No log	6.75	108	0.3476	0.2243	0.3476
No log	6.875	110	0.3538	0.2243	0.3538
No log	7.0	112	0.4262	0.2163	0.4262
No log	7.125	114	0.4667	0.1909	0.4667
No log	7.25	116	0.4562	0.1881	0.4562
No log	7.375	118	0.4023	0.2163	0.4023
No log	7.5	120	0.3664	0.2169	0.3664
No log	7.625	122	0.3487	0.2346	0.3487
No log	7.75	124	0.3715	0.2262	0.3715
No log	7.875	126	0.4267	0.2035	0.4267
No log	8.0	128	0.5126	0.1714	0.5126
No log	8.125	130	0.5465	0.1609	0.5465
No log	8.25	132	0.5097	0.1661	0.5097
No log	8.375	134	0.4320	0.2006	0.4320
No log	8.5	136	0.3786	0.2290	0.3786
No log	8.625	138	0.3721	0.2216	0.3721
No log	8.75	140	0.3867	0.2177	0.3867
No log	8.875	142	0.4094	0.2086	0.4094
No log	9.0	144	0.4304	0.2070	0.4304
No log	9.125	146	0.4412	0.2070	0.4412
No log	9.25	148	0.4352	0.2160	0.4352
No log	9.375	150	0.4157	0.2250	0.4157
No log	9.5	152	0.4008	0.2250	0.4008
No log	9.625	154	0.3937	0.2223	0.3937
No log	9.75	156	0.3953	0.2223	0.3953
No log	9.875	158	0.3987	0.2250	0.3987
No log	10.0	160	0.4006	0.2250	0.4006

Framework versions

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