dinov2-base-finetuned-ct-iq

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

• Loss: 0.0000
• Accuracy: 1.0
• F1: 1.0

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.0265	0.9954	162	0.2460	0.9233	0.9295
0.0185	1.9969	325	0.0023	1.0	1.0
0.1076	2.9985	488	0.0204	0.9939	0.9938
0.1424	4.0	651	0.0001	1.0	1.0
0.0002	4.9954	813	0.0013	1.0	1.0
0.0414	5.9969	976	0.0000	1.0	1.0
0.0003	6.9985	1139	0.0003	1.0	1.0
0.0011	8.0	1302	0.0163	0.9969	0.9969
0.0	8.9954	1464	0.0010	1.0	1.0
0.0	9.9539	1620	0.0000	1.0	1.0

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0
• Datasets 2.21.0
• Tokenizers 0.19.1