vit-msn-small-beta-fia-manually-enhanced-HSV_test_5

This model is a fine-tuned version of facebook/vit-msn-small on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.3267
• Accuracy: 0.9167

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: 1e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 5
• total_train_batch_size: 320
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.25
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.7143	1	1.1106	0.2292
No log	1.4286	2	1.0984	0.2569
No log	2.8571	4	1.0400	0.4097
No log	3.5714	5	0.9960	0.5486
No log	5.0	7	0.8868	0.7292
No log	5.7143	8	0.8263	0.7778
No log	6.4286	9	0.7651	0.8056
0.9808	7.8571	11	0.6521	0.8125
0.9808	8.5714	12	0.6052	0.8125
0.9808	10.0	14	0.5388	0.8125
0.9808	10.7143	15	0.5174	0.8125
0.9808	11.4286	16	0.5032	0.8125
0.9808	12.8571	18	0.5022	0.8125
0.9808	13.5714	19	0.5044	0.8194
0.5431	15.0	21	0.4773	0.8264
0.5431	15.7143	22	0.4439	0.8333
0.5431	16.4286	23	0.4198	0.8403
0.5431	17.8571	25	0.3873	0.8819
0.5431	18.5714	26	0.3730	0.8889
0.5431	20.0	28	0.3774	0.9028
0.5431	20.7143	29	0.3705	0.9097
0.4028	21.4286	30	0.3587	0.9097
0.4028	22.8571	32	0.3662	0.8958
0.4028	23.5714	33	0.3779	0.8681
0.4028	25.0	35	0.4322	0.8264
0.4028	25.7143	36	0.3944	0.8333
0.4028	26.4286	37	0.3585	0.8889
0.4028	27.8571	39	0.3608	0.8889
0.3497	28.5714	40	0.3972	0.8472
0.3497	30.0	42	0.3805	0.8611
0.3497	30.7143	43	0.3611	0.8819
0.3497	31.4286	44	0.3267	0.9167
0.3497	32.8571	46	0.3403	0.9028
0.3497	33.5714	47	0.3751	0.875
0.3497	35.0	49	0.3801	0.8681
0.3278	35.7143	50	0.3499	0.8958
0.3278	36.4286	51	0.3384	0.8958
0.3278	37.8571	53	0.3642	0.8542
0.3278	38.5714	54	0.3997	0.8194
0.3278	40.0	56	0.3843	0.8403
0.3278	40.7143	57	0.3676	0.8681
0.3278	41.4286	58	0.3464	0.9028
0.3334	42.8571	60	0.3618	0.8819
0.3334	43.5714	61	0.4006	0.8194
0.3334	45.0	63	0.4931	0.7639
0.3334	45.7143	64	0.4845	0.7708
0.3334	46.4286	65	0.4485	0.7917
0.3334	47.8571	67	0.3783	0.8472
0.3334	48.5714	68	0.3723	0.8472
0.3334	50.0	70	0.4077	0.8125
0.3334	50.7143	71	0.4381	0.7986
0.3334	51.4286	72	0.4627	0.7847
0.3334	52.8571	74	0.4445	0.7986
0.3334	53.5714	75	0.4141	0.8125
0.3334	55.0	77	0.3489	0.8681
0.3334	55.7143	78	0.3371	0.8958
0.3334	56.4286	79	0.3358	0.8889
0.3105	57.8571	81	0.3539	0.8681
0.3105	58.5714	82	0.3678	0.8542
0.3105	60.0	84	0.3931	0.8264
0.3105	60.7143	85	0.3938	0.8264
0.3105	61.4286	86	0.3897	0.8472
0.3105	62.8571	88	0.3638	0.8611
0.3105	63.5714	89	0.3496	0.875
0.3061	65.0	91	0.3305	0.8958
0.3061	65.7143	92	0.3284	0.9028
0.3061	66.4286	93	0.3284	0.8958
0.3061	67.8571	95	0.3337	0.8958
0.3061	68.5714	96	0.3374	0.8889
0.3061	70.0	98	0.3442	0.875
0.3061	70.7143	99	0.3452	0.875
0.3137	71.4286	100	0.3460	0.875

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.2.0
• Tokenizers 0.19.1