document-spoof-clip

This model is a fine-tuned version of openai/clip-vit-base-patch32 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.0225
• Accuracy: 0.9857

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.8421	4	0.5305	0.8429
No log	1.8947	9	0.1707	0.9286
0.6268	2.9474	14	0.3507	0.8429
0.6268	4.0	19	0.4707	0.8
0.2881	4.8421	23	0.1337	0.9286
0.2881	5.8947	28	0.1293	0.9286
0.2349	6.9474	33	0.0565	0.9714
0.2349	8.0	38	0.0676	0.9571
0.0907	8.8421	42	0.3071	0.9
0.0907	9.8947	47	0.1462	0.9714
0.1203	10.9474	52	0.0761	0.9714
0.1203	12.0	57	0.0808	0.9571
0.0715	12.8421	61	0.0204	0.9857
0.0715	13.8947	66	0.0210	0.9857
0.031	14.9474	71	0.0274	0.9714
0.031	16.0	76	0.0448	0.9857
0.0655	16.8421	80	0.0225	0.9857

Framework versions

• Transformers 4.41.2
• Pytorch 2.1.2
• Datasets 2.19.2
• Tokenizers 0.19.1