ArabicNewSplits6_WithDuplicationsForScore5_FineTuningAraBERT_run2_AugV5_k2_task1_organization

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: 0.8081
• Qwk: 0.6932
• Mse: 0.8081
• Rmse: 0.8989

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: 8
• eval_batch_size: 8
• 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	Rmse
No log	0.1333	2	5.4624	-0.0286	5.4624	2.3372
No log	0.2667	4	3.1895	0.0854	3.1895	1.7859
No log	0.4	6	1.8321	0.1163	1.8321	1.3536
No log	0.5333	8	1.2332	0.2870	1.2332	1.1105
No log	0.6667	10	1.0474	0.3067	1.0474	1.0234
No log	0.8	12	1.1850	0.2120	1.1850	1.0886
No log	0.9333	14	1.2763	0.2596	1.2763	1.1297
No log	1.0667	16	1.0921	0.4320	1.0921	1.0450
No log	1.2	18	1.0071	0.2966	1.0071	1.0035
No log	1.3333	20	0.9948	0.3379	0.9948	0.9974
No log	1.4667	22	0.9277	0.3651	0.9277	0.9632
No log	1.6	24	0.8506	0.3845	0.8506	0.9223
No log	1.7333	26	0.8022	0.4289	0.8022	0.8956
No log	1.8667	28	0.7656	0.4975	0.7656	0.8750
No log	2.0	30	0.7406	0.5603	0.7406	0.8606
No log	2.1333	32	0.7423	0.5516	0.7423	0.8616
No log	2.2667	34	0.7443	0.5783	0.7443	0.8627
No log	2.4	36	0.9241	0.4947	0.9241	0.9613
No log	2.5333	38	1.0845	0.4726	1.0845	1.0414
No log	2.6667	40	0.8550	0.5637	0.8550	0.9247
No log	2.8	42	0.6488	0.6772	0.6488	0.8055
No log	2.9333	44	0.6275	0.6845	0.6275	0.7921
No log	3.0667	46	0.6188	0.6891	0.6188	0.7866
No log	3.2	48	0.6142	0.6748	0.6142	0.7837
No log	3.3333	50	0.6128	0.7009	0.6128	0.7828
No log	3.4667	52	0.6105	0.7170	0.6105	0.7813
No log	3.6	54	0.6784	0.6544	0.6784	0.8237
No log	3.7333	56	0.8618	0.6033	0.8618	0.9283
No log	3.8667	58	0.7382	0.6665	0.7382	0.8592
No log	4.0	60	0.6540	0.6846	0.6540	0.8087
No log	4.1333	62	0.6053	0.7084	0.6053	0.7780
No log	4.2667	64	0.6999	0.6609	0.6999	0.8366
No log	4.4	66	0.6839	0.6667	0.6839	0.8270
No log	4.5333	68	0.6147	0.7122	0.6147	0.7840
No log	4.6667	70	0.6325	0.7021	0.6325	0.7953
No log	4.8	72	0.8044	0.6239	0.8044	0.8969
No log	4.9333	74	0.8076	0.6194	0.8076	0.8987
No log	5.0667	76	0.6800	0.6528	0.6800	0.8246
No log	5.2	78	0.6220	0.6988	0.6220	0.7887
No log	5.3333	80	0.7061	0.7087	0.7061	0.8403
No log	5.4667	82	0.7577	0.6830	0.7577	0.8704
No log	5.6	84	0.7328	0.6940	0.7328	0.8560
No log	5.7333	86	0.7307	0.7193	0.7307	0.8548
No log	5.8667	88	0.7266	0.7177	0.7266	0.8524
No log	6.0	90	0.7419	0.7255	0.7419	0.8613
No log	6.1333	92	0.7896	0.6922	0.7896	0.8886
No log	6.2667	94	0.9077	0.6689	0.9077	0.9527
No log	6.4	96	0.8928	0.6689	0.8928	0.9449
No log	6.5333	98	0.7758	0.6805	0.7758	0.8808
No log	6.6667	100	0.7300	0.7157	0.7300	0.8544
No log	6.8	102	0.7118	0.7161	0.7118	0.8437
No log	6.9333	104	0.6954	0.7382	0.6954	0.8339
No log	7.0667	106	0.6892	0.7195	0.6892	0.8302
No log	7.2	108	0.6713	0.7334	0.6713	0.8194
No log	7.3333	110	0.6761	0.7250	0.6761	0.8223
No log	7.4667	112	0.6816	0.7187	0.6816	0.8256
No log	7.6	114	0.7027	0.7175	0.7027	0.8383
No log	7.7333	116	0.7518	0.7001	0.7518	0.8671
No log	7.8667	118	0.7982	0.6982	0.7982	0.8934
No log	8.0	120	0.8319	0.6947	0.8319	0.9121
No log	8.1333	122	0.8787	0.6930	0.8787	0.9374
No log	8.2667	124	0.8647	0.6951	0.8647	0.9299
No log	8.4	126	0.8419	0.7012	0.8419	0.9176
No log	8.5333	128	0.8374	0.7046	0.8374	0.9151
No log	8.6667	130	0.8294	0.7052	0.8294	0.9107
No log	8.8	132	0.8366	0.7014	0.8366	0.9147
No log	8.9333	134	0.8401	0.7029	0.8401	0.9166
No log	9.0667	136	0.8198	0.6998	0.8198	0.9054
No log	9.2	138	0.8118	0.6842	0.8118	0.9010
No log	9.3333	140	0.8189	0.6842	0.8189	0.9049
No log	9.4667	142	0.8198	0.6932	0.8198	0.9054
No log	9.6	144	0.8224	0.6915	0.8224	0.9069
No log	9.7333	146	0.8125	0.6932	0.8125	0.9014
No log	9.8667	148	0.8082	0.6932	0.8082	0.8990
No log	10.0	150	0.8081	0.6932	0.8081	0.8989

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu118
• Datasets 2.21.0
• Tokenizers 0.19.1