BilalMuftuoglu/beit-base-patch16-224-55-fold5

beit-base-patch16-224-fold5

This model is a fine-tuned version of microsoft/beit-base-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.6919
• Accuracy: 0.8228

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.8571	3	0.6964	0.4937
No log	2.0	7	0.6336	0.6456
0.7161	2.8571	10	0.6712	0.5063
0.7161	4.0	14	0.5884	0.6835
0.7161	4.8571	17	0.6325	0.5570
0.6466	6.0	21	0.6307	0.6076
0.6466	6.8571	24	0.5518	0.6835
0.6466	8.0	28	0.6856	0.6456
0.5604	8.8571	31	0.5678	0.7089
0.5604	10.0	35	0.5764	0.6329
0.5604	10.8571	38	0.6124	0.6456
0.4941	12.0	42	1.1315	0.5316
0.4941	12.8571	45	0.5624	0.6835
0.4941	14.0	49	0.6351	0.6962
0.4516	14.8571	52	0.6811	0.6835
0.4516	16.0	56	0.7385	0.6456
0.4516	16.8571	59	0.6123	0.7215
0.409	18.0	63	0.5877	0.6835
0.409	18.8571	66	0.6250	0.6962
0.3439	20.0	70	0.6220	0.7342
0.3439	20.8571	73	0.7250	0.7089
0.3439	22.0	77	0.6107	0.7342
0.3268	22.8571	80	0.4999	0.7975
0.3268	24.0	84	0.5325	0.7468
0.3268	24.8571	87	0.7209	0.7342
0.2941	26.0	91	0.5557	0.7722
0.2941	26.8571	94	0.6655	0.7595
0.2941	28.0	98	1.0775	0.6962
0.284	28.8571	101	0.6817	0.7595
0.284	30.0	105	0.9235	0.6835
0.284	30.8571	108	0.6587	0.7595
0.3134	32.0	112	0.7086	0.7468
0.3134	32.8571	115	0.6895	0.7468
0.3134	34.0	119	0.6418	0.7722
0.2266	34.8571	122	0.7007	0.7848
0.2266	36.0	126	0.6919	0.8228
0.2266	36.8571	129	0.7562	0.7342
0.2249	38.0	133	0.6775	0.7722
0.2249	38.8571	136	0.7787	0.7595
0.2181	40.0	140	0.7932	0.7722
0.2181	40.8571	143	0.9334	0.7595
0.2181	42.0	147	0.8224	0.7468
0.186	42.8571	150	0.8444	0.7722
0.186	44.0	154	1.0350	0.7722
0.186	44.8571	157	0.8386	0.7722
0.1882	46.0	161	0.8384	0.7722
0.1882	46.8571	164	0.7905	0.7595
0.1882	48.0	168	0.7184	0.7722
0.1649	48.8571	171	0.8119	0.7595
0.1649	50.0	175	0.7172	0.7722
0.1649	50.8571	178	0.9249	0.7722
0.1847	52.0	182	0.8069	0.7722
0.1847	52.8571	185	0.8753	0.7722
0.1847	54.0	189	0.8822	0.7722
0.1606	54.8571	192	0.7481	0.7975
0.1606	56.0	196	0.7300	0.7722
0.1606	56.8571	199	0.7325	0.8101
0.142	58.0	203	0.7088	0.8228
0.142	58.8571	206	0.6900	0.8228
0.1546	60.0	210	0.8303	0.7722
0.1546	60.8571	213	0.8218	0.7975
0.1546	62.0	217	0.9628	0.7848
0.1608	62.8571	220	0.9826	0.7848
0.1608	64.0	224	0.7604	0.7595
0.1608	64.8571	227	0.8165	0.7722
0.1418	66.0	231	0.8280	0.7722
0.1418	66.8571	234	0.9254	0.7848
0.1418	68.0	238	0.8439	0.7848
0.1505	68.8571	241	0.8096	0.7975
0.1505	70.0	245	0.9471	0.7595
0.1505	70.8571	248	0.9028	0.7595
0.1211	72.0	252	0.7744	0.7975
0.1211	72.8571	255	0.7756	0.8101
0.1211	74.0	259	0.7904	0.8101
0.1248	74.8571	262	0.8036	0.8101
0.1248	76.0	266	0.8592	0.7848
0.1248	76.8571	269	0.8946	0.7848
0.122	78.0	273	0.8576	0.7848
0.122	78.8571	276	0.9037	0.7975
0.1129	80.0	280	0.9982	0.7848
0.1129	80.8571	283	1.0404	0.7722
0.1129	82.0	287	0.9969	0.7975
0.1136	82.8571	290	0.9586	0.7975
0.1136	84.0	294	0.9200	0.7848
0.1136	84.8571	297	0.9063	0.7722
0.1233	85.7143	300	0.9050	0.7722

Framework versions

• Transformers 4.40.2
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1