outputs

This model is a fine-tuned version of facebook/dinov2-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0742
• Precision: 0.9306
• Recall: 0.8969
• F1: 0.9135

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-06
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 8
• 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: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1
0.5948	0.98	39	0.4487	0.1103	0.0658	0.0824
0.2211	1.98	79	0.2079	0.8179	0.5614	0.6658
0.1241	2.98	119	0.1378	0.8880	0.7390	0.8067
0.0954	3.99	159	0.1117	0.8916	0.8114	0.8496
0.0801	4.99	199	0.0980	0.9167	0.8322	0.8724
0.0716	5.99	239	0.0875	0.9245	0.8596	0.8909
0.0641	7.0	279	0.0871	0.9231	0.8421	0.8807
0.0615	8.0	319	0.0804	0.9318	0.8838	0.9071
0.056	8.98	358	0.0793	0.9257	0.8882	0.9065
0.0541	9.98	398	0.0761	0.9335	0.8925	0.9126
0.0532	10.98	438	0.0767	0.9339	0.8827	0.9076
0.053	11.99	478	0.0758	0.9312	0.8904	0.9103
0.048	12.99	518	0.0743	0.9324	0.8925	0.9120
0.047	13.99	558	0.0750	0.9303	0.8925	0.9110
0.0476	14.67	585	0.0742	0.9306	0.8969	0.9135

Framework versions

• Transformers 4.37.0
• Pytorch 1.13.1+cu117
• Datasets 2.18.0
• Tokenizers 0.15.2