pcernuta/klobasa-ni-klobasa

convnext-tiny-224-klobasaniklobasa

This model is a fine-tuned version of facebook/convnext-tiny-224 on a small dataset of klobasa images scraped from the internets. It achieves the following results on the evaluation set:

• Loss: 0.4401
• Accuracy: 0.8958

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: 4e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• 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
0.7062	1.0	48	0.7116	0.8438
0.4831	2.0	96	0.5968	0.8333
0.2429	3.0	144	0.5384	0.8542
0.2292	4.0	192	0.4995	0.8438
0.1549	5.0	240	0.4508	0.8854
0.1682	6.0	288	0.4401	0.8958
0.1736	7.0	336	0.4440	0.8958
0.0633	8.0	384	0.4406	0.8958
0.0689	9.0	432	0.4371	0.8958
0.0558	10.0	480	0.4334	0.8958

Framework versions

• Transformers 4.25.1
• Pytorch 1.13.0+cu116
• Datasets 2.8.0
• Tokenizers 0.13.2