ArabicNewSplits6_WithDuplicationsForScore5_FineTuningAraBERT_run3_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: 0.7546
• Qwk: 0.7354
• Mse: 0.7546
• Rmse: 0.8687

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	5.1824	-0.0323	5.1824	2.2765
No log	0.2667	4	3.6030	0.0707	3.6030	1.8982
No log	0.4	6	2.0550	0.1300	2.0550	1.4335
No log	0.5333	8	1.8337	0.0019	1.8337	1.3541
No log	0.6667	10	1.6659	0.0361	1.6659	1.2907
No log	0.8	12	1.3265	0.1447	1.3265	1.1517
No log	0.9333	14	1.2051	0.1688	1.2051	1.0977
No log	1.0667	16	1.2359	0.1767	1.2359	1.1117
No log	1.2	18	1.0991	0.3203	1.0991	1.0484
No log	1.3333	20	0.9935	0.4469	0.9935	0.9967
No log	1.4667	22	1.0461	0.4449	1.0461	1.0228
No log	1.6	24	1.0640	0.4361	1.0640	1.0315
No log	1.7333	26	0.8554	0.5456	0.8554	0.9249
No log	1.8667	28	0.8140	0.5341	0.8140	0.9022
No log	2.0	30	0.7995	0.5140	0.7995	0.8942
No log	2.1333	32	0.8216	0.4881	0.8216	0.9064
No log	2.2667	34	0.8021	0.5192	0.8021	0.8956
No log	2.4	36	0.8038	0.5480	0.8038	0.8966
No log	2.5333	38	0.8426	0.5590	0.8426	0.9179
No log	2.6667	40	0.9434	0.5055	0.9434	0.9713
No log	2.8	42	0.8122	0.6420	0.8122	0.9012
No log	2.9333	44	0.7790	0.6365	0.7790	0.8826
No log	3.0667	46	0.7981	0.6084	0.7981	0.8934
No log	3.2	48	0.7567	0.6459	0.7567	0.8699
No log	3.3333	50	0.8096	0.6704	0.8096	0.8998
No log	3.4667	52	0.8975	0.6454	0.8975	0.9474
No log	3.6	54	0.8738	0.6486	0.8738	0.9348
No log	3.7333	56	0.7492	0.6870	0.7492	0.8656
No log	3.8667	58	0.7177	0.6609	0.7177	0.8472
No log	4.0	60	0.7220	0.6773	0.7220	0.8497
No log	4.1333	62	0.7891	0.6835	0.7891	0.8883
No log	4.2667	64	0.7116	0.6963	0.7116	0.8435
No log	4.4	66	0.7464	0.6928	0.7464	0.8640
No log	4.5333	68	0.7899	0.6833	0.7899	0.8887
No log	4.6667	70	0.6895	0.7051	0.6895	0.8304
No log	4.8	72	0.7664	0.7202	0.7664	0.8754
No log	4.9333	74	1.1343	0.5577	1.1343	1.0650
No log	5.0667	76	1.2956	0.4826	1.2956	1.1382
No log	5.2	78	1.1725	0.5325	1.1725	1.0828
No log	5.3333	80	0.8672	0.7006	0.8672	0.9312
No log	5.4667	82	0.6691	0.7084	0.6691	0.8180
No log	5.6	84	0.6726	0.6943	0.6726	0.8201
No log	5.7333	86	0.6693	0.6942	0.6693	0.8181
No log	5.8667	88	0.6796	0.7100	0.6796	0.8244
No log	6.0	90	0.7565	0.7273	0.7565	0.8698
No log	6.1333	92	0.9236	0.6631	0.9236	0.9611
No log	6.2667	94	0.9062	0.6631	0.9062	0.9519
No log	6.4	96	0.7624	0.7328	0.7624	0.8732
No log	6.5333	98	0.6519	0.7273	0.6519	0.8074
No log	6.6667	100	0.6738	0.6890	0.6738	0.8209
No log	6.8	102	0.6940	0.6872	0.6940	0.8331
No log	6.9333	104	0.6509	0.6998	0.6509	0.8068
No log	7.0667	106	0.6345	0.7175	0.6345	0.7965
No log	7.2	108	0.6947	0.7391	0.6947	0.8335
No log	7.3333	110	0.8361	0.6968	0.8361	0.9144
No log	7.4667	112	0.8875	0.6820	0.8875	0.9420
No log	7.6	114	0.9409	0.6397	0.9409	0.9700
No log	7.7333	116	0.8663	0.6790	0.8663	0.9308
No log	7.8667	118	0.7662	0.7349	0.7662	0.8753
No log	8.0	120	0.6960	0.7485	0.6960	0.8343
No log	8.1333	122	0.6706	0.7539	0.6706	0.8189
No log	8.2667	124	0.6755	0.7576	0.6755	0.8219
No log	8.4	126	0.7013	0.7488	0.7013	0.8374
No log	8.5333	128	0.7386	0.7298	0.7386	0.8594
No log	8.6667	130	0.7691	0.7328	0.7691	0.8770
No log	8.8	132	0.8195	0.7084	0.8195	0.9053
No log	8.9333	134	0.8380	0.7047	0.8380	0.9154
No log	9.0667	136	0.8231	0.7176	0.8231	0.9073
No log	9.2	138	0.7941	0.7363	0.7941	0.8911
No log	9.3333	140	0.7670	0.7389	0.7670	0.8758
No log	9.4667	142	0.7538	0.7354	0.7538	0.8682
No log	9.6	144	0.7484	0.7420	0.7484	0.8651
No log	9.7333	146	0.7527	0.7319	0.7527	0.8676
No log	9.8667	148	0.7536	0.7319	0.7536	0.8681
No log	10.0	150	0.7546	0.7354	0.7546	0.8687

Framework versions

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