ktp-kk-crop

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: 0.0312
• Accuracy: 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: 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: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.8696	5	0.5871	0.7
No log	1.9130	11	0.0729	0.9667
0.7676	2.9565	17	0.1986	0.9
0.7676	4.0	23	0.1610	0.9
0.7676	4.8696	28	0.0644	0.9667
0.2441	5.9130	34	0.2016	0.9
0.2441	6.9565	40	0.1530	0.9
0.1751	8.0	46	0.0412	1.0
0.1751	8.8696	51	0.0301	1.0
0.1751	9.9130	57	0.0495	0.9667
0.1156	10.9565	63	0.0283	1.0
0.1156	12.0	69	0.0214	1.0
0.1156	12.8696	74	0.1014	0.9667
0.1238	13.9130	80	0.0538	1.0
0.1238	14.9565	86	0.0477	1.0
0.1064	16.0	92	0.0105	1.0
0.1064	16.8696	97	0.0389	0.9667
0.1064	17.9130	103	0.0120	1.0
0.0862	18.9565	109	0.0183	1.0
0.0862	20.0	115	0.0259	1.0
0.0345	20.8696	120	0.0272	1.0
0.0345	21.7391	125	0.0312	1.0

Framework versions

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