sudo-s/exper_batch_8_e8 - AI Model Zoo

exper_batch_8_e8

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the sudo-s/herbier_mesuem1 dataset. It achieves the following results on the evaluation set:

Loss: 0.4608
Accuracy: 0.9052

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: 0.0002
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: 8
mixed_precision_training: Apex, opt level O1

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
4.2202	0.08	100	4.1245	0.1237
3.467	0.16	200	3.5622	0.2143
3.3469	0.23	300	3.1688	0.2675
2.8086	0.31	400	2.8965	0.3034
2.6291	0.39	500	2.5858	0.4025
2.2382	0.47	600	2.2908	0.4133
1.9259	0.55	700	2.2007	0.4676
1.8088	0.63	800	2.0419	0.4742
1.9462	0.7	900	1.6793	0.5578
1.5392	0.78	1000	1.5460	0.6079
1.561	0.86	1100	1.5793	0.5690
1.2135	0.94	1200	1.4663	0.5929
1.0725	1.02	1300	1.2974	0.6534
0.8696	1.1	1400	1.2406	0.6569
0.8758	1.17	1500	1.2127	0.6623
1.1737	1.25	1600	1.2243	0.6550
0.8242	1.33	1700	1.1371	0.6735
1.0141	1.41	1800	1.0536	0.7024
0.9855	1.49	1900	0.9885	0.7205
0.805	1.57	2000	0.9048	0.7479
0.7207	1.64	2100	0.8842	0.7490
0.7101	1.72	2200	0.8954	0.7436
0.5946	1.8	2300	0.9174	0.7386
0.6937	1.88	2400	0.7818	0.7760
0.5593	1.96	2500	0.7449	0.7934
0.4139	2.04	2600	0.7787	0.7830
0.2929	2.11	2700	0.7122	0.7945
0.4159	2.19	2800	0.7446	0.7907
0.4079	2.27	2900	0.7354	0.7938
0.516	2.35	3000	0.7499	0.8007
0.2728	2.43	3100	0.6851	0.8061
0.4159	2.51	3200	0.7258	0.7999
0.3396	2.58	3300	0.7455	0.7972
0.1918	2.66	3400	0.6793	0.8119
0.1228	2.74	3500	0.6696	0.8134
0.2671	2.82	3600	0.6306	0.8285
0.4986	2.9	3700	0.6111	0.8296
0.3699	2.98	3800	0.5600	0.8508
0.0444	3.05	3900	0.6021	0.8331
0.1489	3.13	4000	0.5599	0.8516
0.15	3.21	4100	0.6377	0.8365
0.2535	3.29	4200	0.5752	0.8543
0.2679	3.37	4300	0.5677	0.8608
0.0989	3.45	4400	0.6325	0.8396
0.0825	3.52	4500	0.5979	0.8524
0.0427	3.6	4600	0.5903	0.8516
0.1806	3.68	4700	0.5323	0.8628
0.2672	3.76	4800	0.5688	0.8604
0.2674	3.84	4900	0.5369	0.8635
0.2185	3.92	5000	0.4743	0.8820
0.2195	3.99	5100	0.5340	0.8709
0.0049	4.07	5200	0.5883	0.8608
0.0204	4.15	5300	0.6102	0.8539
0.0652	4.23	5400	0.5659	0.8670
0.028	4.31	5500	0.4916	0.8840
0.0423	4.39	5600	0.5706	0.8736
0.0087	4.46	5700	0.5653	0.8697
0.0964	4.54	5800	0.5423	0.8755
0.0841	4.62	5900	0.5160	0.8743
0.0945	4.7	6000	0.5532	0.8697
0.0311	4.78	6100	0.4947	0.8867
0.0423	4.86	6200	0.5063	0.8843
0.1348	4.93	6300	0.5619	0.8743
0.049	5.01	6400	0.5800	0.8732
0.0053	5.09	6500	0.5499	0.8770
0.0234	5.17	6600	0.5102	0.8874
0.0192	5.25	6700	0.5447	0.8836
0.0029	5.32	6800	0.4787	0.8936
0.0249	5.4	6900	0.5232	0.8870
0.0671	5.48	7000	0.4766	0.8975
0.0056	5.56	7100	0.5136	0.8894
0.003	5.64	7200	0.5085	0.8882
0.0015	5.72	7300	0.4832	0.8971
0.0014	5.79	7400	0.4648	0.8998
0.0065	5.87	7500	0.4739	0.8978
0.0011	5.95	7600	0.5349	0.8867
0.0021	6.03	7700	0.5460	0.8847
0.0012	6.11	7800	0.5309	0.8890
0.0011	6.19	7900	0.4852	0.8998
0.0093	6.26	8000	0.4751	0.8998
0.003	6.34	8100	0.4934	0.8963
0.0027	6.42	8200	0.4882	0.9029
0.0009	6.5	8300	0.4806	0.9021
0.0009	6.58	8400	0.4974	0.9029
0.0009	6.66	8500	0.4748	0.9075
0.0008	6.73	8600	0.4723	0.9094
0.001	6.81	8700	0.4692	0.9098
0.0007	6.89	8800	0.4726	0.9075
0.0011	6.97	8900	0.4686	0.9067
0.0006	7.05	9000	0.4653	0.9056
0.0006	7.13	9100	0.4755	0.9029
0.0007	7.2	9200	0.4633	0.9036
0.0067	7.28	9300	0.4611	0.9036
0.0007	7.36	9400	0.4608	0.9052
0.0007	7.44	9500	0.4623	0.9044
0.0005	7.52	9600	0.4621	0.9056
0.0005	7.6	9700	0.4615	0.9056
0.0005	7.67	9800	0.4612	0.9059
0.0005	7.75	9900	0.4626	0.9075
0.0004	7.83	10000	0.4626	0.9075
0.0005	7.91	10100	0.4626	0.9075
0.0006	7.99	10200	0.4626	0.9079

Framework versions

Transformers 4.19.4
Pytorch 1.5.1
Datasets 2.3.2
Tokenizers 0.12.1