al-css/faces_classification

faces_clasification_alcss

This model is a fine-tuned version of lokeshk/Face-Recognition-NM on the private_faces_dataset dataset. It achieves the following results on the evaluation set:

• Loss: 2.2189
• Accuracy: 0.5481

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: 0.0001
• 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	Accuracy
4.6929	0.0962	10	4.5349	0.0913
4.4502	0.1923	20	4.3568	0.1010
4.3178	0.2885	30	4.2414	0.0865
4.2083	0.3846	40	4.1396	0.1731
4.1225	0.4808	50	4.0550	0.1538
3.908	0.5769	60	3.9737	0.1683
3.9502	0.6731	70	3.9053	0.1683
3.8725	0.7692	80	3.8432	0.1731
3.7944	0.8654	90	3.7840	0.2115
3.7648	0.9615	100	3.7644	0.1971
3.7211	1.0577	110	3.7172	0.1971
3.5759	1.1538	120	3.6622	0.25
3.6082	1.25	130	3.6080	0.2644
3.4272	1.3462	140	3.5738	0.2837
3.3349	1.4423	150	3.5526	0.2740
3.189	1.5385	160	3.5162	0.2837
3.2959	1.6346	170	3.4639	0.2933
3.1273	1.7308	180	3.4308	0.3365
3.1621	1.8269	190	3.3436	0.3221
3.1053	1.9231	200	3.2973	0.3702
3.0641	2.0192	210	3.2637	0.3462
2.6593	2.1154	220	3.2149	0.3798
2.5709	2.2115	230	3.1658	0.3846
2.5412	2.3077	240	3.1359	0.3702
2.5775	2.4038	250	3.1104	0.3654
2.6372	2.5	260	3.1027	0.3798
2.6732	2.5962	270	3.0252	0.4231
2.6384	2.6923	280	3.0202	0.3942
2.4607	2.7885	290	2.9785	0.4135
2.5519	2.8846	300	2.9470	0.4327
2.2381	2.9808	310	2.9402	0.4231
2.1999	3.0769	320	2.9074	0.4519
2.179	3.1731	330	2.8780	0.4567
2.1427	3.2692	340	2.8331	0.4423
2.1335	3.3654	350	2.8051	0.4760
1.7641	3.4615	360	2.7798	0.4712
1.9687	3.5577	370	2.7607	0.4808
1.8046	3.6538	380	2.7381	0.4760
1.944	3.75	390	2.7244	0.4856
1.7403	3.8462	400	2.6899	0.4567
1.7732	3.9423	410	2.6656	0.4808
1.4105	4.0385	420	2.6526	0.4760
1.377	4.1346	430	2.6448	0.4904
1.5767	4.2308	440	2.5933	0.4663
1.3826	4.3269	450	2.5832	0.5
1.6504	4.4231	460	2.5573	0.5192
1.5579	4.5192	470	2.5666	0.5048
1.2466	4.6154	480	2.5197	0.5144
1.32	4.7115	490	2.5145	0.5240
1.5286	4.8077	500	2.4909	0.5192
1.394	4.9038	510	2.4882	0.5192
1.3982	5.0	520	2.4616	0.5192
1.1167	5.0962	530	2.4569	0.5096
1.3562	5.1923	540	2.4559	0.5192
1.0018	5.2885	550	2.4438	0.5192
1.2367	5.3846	560	2.4204	0.5192
1.1748	5.4808	570	2.4112	0.5337
1.022	5.5769	580	2.4130	0.5385
1.0954	5.6731	590	2.3992	0.5192
0.9759	5.7692	600	2.3683	0.5240
1.0327	5.8654	610	2.3577	0.5144
1.1167	5.9615	620	2.3547	0.5144
0.8077	6.0577	630	2.3452	0.5385
0.95	6.1538	640	2.3486	0.5433
0.7993	6.25	650	2.3428	0.5529
0.923	6.3462	660	2.3279	0.5337
0.7566	6.4423	670	2.3176	0.5385
0.8834	6.5385	680	2.3201	0.5288
0.9337	6.6346	690	2.3064	0.5529
0.7596	6.7308	700	2.3063	0.5288
0.973	6.8269	710	2.2847	0.5337
1.0212	6.9231	720	2.3006	0.5433
0.8315	7.0192	730	2.2813	0.5385
0.814	7.1154	740	2.2751	0.5481
0.7658	7.2115	750	2.2754	0.5433
0.5956	7.3077	760	2.2781	0.5433
0.8864	7.4038	770	2.2602	0.5529
0.7181	7.5	780	2.2568	0.5721
0.643	7.5962	790	2.2586	0.5481
0.7107	7.6923	800	2.2520	0.5337
0.6078	7.7885	810	2.2423	0.5385
0.873	7.8846	820	2.2478	0.5337
0.7018	7.9808	830	2.2357	0.5577
0.7177	8.0769	840	2.2340	0.5481
0.6543	8.1731	850	2.2337	0.5481
0.7563	8.2692	860	2.2259	0.5529
0.549	8.3654	870	2.2310	0.5529
0.6981	8.4615	880	2.2381	0.5529
0.6172	8.5577	890	2.2285	0.5481
0.587	8.6538	900	2.2165	0.5481
0.613	8.75	910	2.2215	0.5481
0.5558	8.8462	920	2.2274	0.5529
0.6892	8.9423	930	2.2221	0.5433
0.5876	9.0385	940	2.2209	0.5481
0.5845	9.1346	950	2.2211	0.5481
0.645	9.2308	960	2.2208	0.5481
0.5619	9.3269	970	2.2204	0.5481
0.5371	9.4231	980	2.2219	0.5481
0.5048	9.5192	990	2.2203	0.5481
0.6007	9.6154	1000	2.2192	0.5481
0.5706	9.7115	1010	2.2187	0.5481
0.571	9.8077	1020	2.2189	0.5481
0.6692	9.9038	1030	2.2191	0.5481
0.5411	10.0	1040	2.2189	0.5481

Framework versions

• Transformers 4.44.1
• Pytorch 2.4.0+cu118
• Datasets 2.21.0
• Tokenizers 0.19.1