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: 0.1223
Accuracy: 0.9865

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	0.96	6	0.8954	0.5270
0.7112	1.92	12	0.6729	0.5405
0.7112	2.88	18	0.6407	0.7297
0.4413	4.0	25	0.1279	0.9459
0.0935	4.96	31	0.1436	0.9730
0.0935	5.92	37	0.0021	1.0
0.0697	6.88	43	0.2862	0.9459
0.161	8.0	50	0.0843	0.9595
0.161	8.96	56	0.2255	0.9459
0.0061	9.92	62	0.4678	0.9054
0.0061	10.88	68	0.3299	0.9189
0.0309	12.0	75	0.5189	0.9189
0.0025	12.96	81	0.0850	0.9865
0.0025	13.92	87	0.0720	0.9865
0.0042	14.88	93	0.0745	0.9865
0.0002	16.0	100	0.0869	0.9865
0.0002	16.96	106	0.0895	0.9865
0.0001	17.92	112	0.1127	0.9865
0.0001	18.88	118	0.1219	0.9865
0.0	19.2	120	0.1223	0.9865

Framework versions

Transformers 4.41.2
Pytorch 2.1.2
Datasets 2.19.2
Tokenizers 0.19.1

habibi26/ktp-crop-clip

ktp-crop-clip

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

habibi26/
ktp-crop-clip