Aina-large-2024_10_23-batch-size32_freeze_monolabel

This model is a fine-tuned version of facebook/dinov2-large on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6806
• F1 Micro: 0.7614
• F1 Macro: 0.4269
• Accuracy: 0.7614
• Learning Rate: 0.0000

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: 0.001
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 150
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1 Micro	F1 Macro	Accuracy	Rate
0.9658	1.0	3312	0.8468	0.7179	0.2217	0.7179	0.001
0.9257	2.0	6624	0.8172	0.7247	0.3043	0.7247	0.001
0.9202	3.0	9936	0.8048	0.7260	0.3035	0.7260	0.001
0.8905	4.0	13248	0.7947	0.7285	0.3109	0.7285	0.001
0.907	5.0	16560	0.7822	0.7309	0.3046	0.7309	0.001
0.8925	6.0	19872	0.7838	0.7345	0.3159	0.7345	0.001
0.8922	7.0	23184	0.7931	0.7357	0.3244	0.7357	0.001
0.883	8.0	26496	0.7688	0.7354	0.3241	0.7354	0.001
0.8697	9.0	29808	0.7635	0.7377	0.3242	0.7377	0.001
0.8782	10.0	33120	0.7689	0.7373	0.3327	0.7373	0.001
0.8869	11.0	36432	0.7676	0.7350	0.3337	0.7350	0.001
0.8791	12.0	39744	0.7640	0.7369	0.3409	0.7369	0.001
0.9017	13.0	43056	0.7674	0.7337	0.3400	0.7337	0.001
0.8753	14.0	46368	0.7586	0.7381	0.3271	0.7381	0.001
0.872	15.0	49680	0.7658	0.7373	0.3229	0.7373	0.001
0.8672	16.0	52992	0.8086	0.7389	0.3353	0.7389	0.001
0.8678	17.0	56304	0.7629	0.7390	0.3359	0.7390	0.001
0.8875	18.0	59616	0.7615	0.7365	0.3353	0.7365	0.001
0.8645	19.0	62928	0.7682	0.7387	0.3450	0.7387	0.001
0.881	20.0	66240	0.7559	0.7406	0.3411	0.7406	0.001
0.8927	21.0	69552	0.7755	0.7349	0.3408	0.7349	0.001
0.8704	22.0	72864	0.7674	0.7344	0.3233	0.7344	0.001
0.8711	23.0	76176	0.7695	0.7340	0.3139	0.7340	0.001
0.8722	24.0	79488	0.7538	0.7400	0.3338	0.7400	0.001
0.884	25.0	82800	0.7643	0.7352	0.3480	0.7352	0.001
0.8661	26.0	86112	0.7568	0.7388	0.3272	0.7388	0.001
0.8847	27.0	89424	0.7665	0.7371	0.3427	0.7371	0.001
0.8749	28.0	92736	0.7592	0.7385	0.3129	0.7385	0.001
0.8782	29.0	96048	0.7544	0.7402	0.3420	0.7402	0.001
0.882	30.0	99360	0.7549	0.7412	0.3503	0.7412	0.001
0.8481	31.0	102672	0.7332	0.7457	0.3602	0.7457	0.0001
0.8329	32.0	105984	0.7296	0.7456	0.3696	0.7456	0.0001
0.817	33.0	109296	0.7270	0.7467	0.3749	0.7467	0.0001
0.8173	34.0	112608	0.7234	0.7471	0.3683	0.7471	0.0001
0.8221	35.0	115920	0.7187	0.7492	0.3795	0.7492	0.0001
0.8085	36.0	119232	0.7215	0.7484	0.3758	0.7484	0.0001
0.8113	37.0	122544	0.7180	0.7505	0.3767	0.7505	0.0001
0.802	38.0	125856	0.7137	0.7502	0.3861	0.7502	0.0001
0.8042	39.0	129168	0.7125	0.7514	0.3868	0.7514	0.0001
0.7976	40.0	132480	0.7126	0.7499	0.3844	0.7499	0.0001
0.7963	41.0	135792	0.7112	0.7516	0.3905	0.7516	0.0001
0.8054	42.0	139104	0.7116	0.7511	0.3926	0.7511	0.0001
0.8119	43.0	142416	0.7098	0.7516	0.3901	0.7516	0.0001
0.8009	44.0	145728	0.7102	0.7507	0.3897	0.7507	0.0001
0.7929	45.0	149040	0.7100	0.7517	0.3883	0.7517	0.0001
0.8079	46.0	152352	0.7068	0.7510	0.3912	0.7510	0.0001
0.8053	47.0	155664	0.7074	0.7510	0.3888	0.7510	0.0001
0.7965	48.0	158976	0.7095	0.7508	0.3890	0.7508	0.0001
0.8043	49.0	162288	0.7090	0.7509	0.3935	0.7509	0.0001
0.7861	50.0	165600	0.7080	0.7512	0.4026	0.7512	0.0001
0.7917	51.0	168912	0.7062	0.7514	0.3942	0.7514	0.0001
0.7909	52.0	172224	0.7049	0.7526	0.3971	0.7526	0.0001
0.7886	53.0	175536	0.7044	0.7526	0.4017	0.7526	0.0001
0.7834	54.0	178848	0.7028	0.7524	0.3992	0.7524	0.0001
0.7991	55.0	182160	0.7029	0.7527	0.3966	0.7527	0.0001
0.7875	56.0	185472	0.7026	0.7533	0.4011	0.7533	0.0001
0.7868	57.0	188784	0.7029	0.7525	0.4056	0.7525	0.0001
0.7837	58.0	192096	0.7021	0.7536	0.4020	0.7536	0.0001
0.7834	59.0	195408	0.7011	0.7534	0.4049	0.7534	0.0001
0.7893	60.0	198720	0.7019	0.7530	0.4029	0.7530	0.0001
0.7824	61.0	202032	0.7023	0.7519	0.3995	0.7519	0.0001
0.789	62.0	205344	0.7038	0.7525	0.4041	0.7525	0.0001
0.7778	63.0	208656	0.7003	0.7535	0.4038	0.7535	0.0001
0.7719	64.0	211968	0.6997	0.7526	0.3982	0.7526	0.0001
0.7909	65.0	215280	0.7074	0.7515	0.3997	0.7515	0.0001
0.7854	66.0	218592	0.7018	0.7526	0.3940	0.7526	0.0001
0.7746	67.0	221904	0.7023	0.7543	0.4000	0.7543	0.0001
0.7905	68.0	225216	0.6975	0.7541	0.4063	0.7541	0.0001
0.7824	69.0	228528	0.6994	0.7538	0.4072	0.7538	0.0001
0.7795	70.0	231840	0.6969	0.7557	0.4094	0.7557	0.0001
0.7763	71.0	235152	0.6969	0.7564	0.4085	0.7564	0.0001
0.7723	72.0	238464	0.6987	0.7531	0.4090	0.7531	0.0001
0.7914	73.0	241776	0.6945	0.7556	0.4203	0.7556	0.0001
0.7658	74.0	245088	0.6951	0.7544	0.4117	0.7544	0.0001
0.7803	75.0	248400	0.6989	0.7548	0.4104	0.7548	0.0001
0.7772	76.0	251712	0.6997	0.7536	0.4037	0.7536	0.0001
0.7813	77.0	255024	0.6986	0.7535	0.4092	0.7535	0.0001
0.7938	78.0	258336	0.6982	0.7530	0.4084	0.7530	0.0001
0.776	79.0	261648	0.6958	0.7545	0.4055	0.7545	0.0001
0.7613	80.0	264960	0.6934	0.7548	0.4061	0.7548	1e-05
0.7647	81.0	268272	0.6922	0.7560	0.4108	0.7560	1e-05
0.7842	82.0	271584	0.6933	0.7543	0.4069	0.7543	1e-05
0.7689	83.0	274896	0.6953	0.7535	0.4068	0.7535	1e-05
0.7674	84.0	278208	0.6913	0.7570	0.4140	0.7570	1e-05
0.7607	85.0	281520	0.6911	0.7564	0.4117	0.7564	1e-05
0.7744	86.0	284832	0.6916	0.7563	0.4128	0.7563	1e-05
0.7639	87.0	288144	0.6929	0.7550	0.4089	0.7550	1e-05
0.7515	88.0	291456	0.6904	0.7565	0.4210	0.7565	1e-05
0.7529	89.0	294768	0.6912	0.7554	0.4082	0.7554	1e-05
0.7575	90.0	298080	0.6931	0.7557	0.4102	0.7557	1e-05
0.7715	91.0	301392	0.6912	0.7555	0.4130	0.7555	1e-05
0.7512	92.0	304704	0.6950	0.7534	0.4113	0.7534	1e-05
0.7514	93.0	308016	0.6945	0.7539	0.4075	0.7539	1e-05
0.7529	94.0	311328	0.6904	0.7564	0.4140	0.7564	1e-05
0.7731	95.0	314640	0.6919	0.7555	0.4121	0.7555	0.0000
0.7561	96.0	317952	0.6894	0.7563	0.4092	0.7563	0.0000
0.7702	97.0	321264	0.6900	0.7565	0.4131	0.7565	0.0000
0.7506	98.0	324576	0.6900	0.7566	0.4136	0.7566	0.0000
0.7512	99.0	327888	0.6909	0.7564	0.4168	0.7564	0.0000
0.7694	100.0	331200	0.6912	0.7562	0.4155	0.7562	0.0000
0.7487	101.0	334512	0.6904	0.7550	0.4158	0.7550	0.0000
0.7543	102.0	337824	0.6890	0.7570	0.4175	0.7570	0.0000
0.7743	103.0	341136	0.6923	0.7546	0.4137	0.7546	0.0000
0.757	104.0	344448	0.6912	0.7560	0.4183	0.7560	0.0000
0.7631	105.0	347760	0.6899	0.7561	0.4088	0.7561	0.0000
0.755	106.0	351072	0.6912	0.7556	0.4102	0.7556	0.0000
0.7545	107.0	354384	0.6898	0.7573	0.4107	0.7573	0.0000
0.7533	108.0	357696	0.6910	0.7538	0.4114	0.7538	0.0000
0.7725	109.0	361008	0.6899	0.7565	0.4134	0.7565	0.0000
0.7544	110.0	364320	0.6922	0.7555	0.4110	0.7555	0.0000
0.758	111.0	367632	0.6901	0.7559	0.4141	0.7559	0.0000
0.7674	112.0	370944	0.6903	0.7560	0.4127	0.7560	0.0000

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1