resnet-50-finetuned-omars5

This model is a fine-tuned version of microsoft/resnet-50 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.5844
• Accuracy: 0.8845

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: 0.0005
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.3431	0.99	92	1.2810	0.5836
1.0465	2.0	185	0.8740	0.8176
0.8755	2.99	277	0.6467	0.7994
0.7459	4.0	370	0.5379	0.8480
0.7983	4.99	462	0.4385	0.8207
0.7692	6.0	555	0.5795	0.7842
0.5158	6.99	647	0.4936	0.8207
0.625	8.0	740	0.5316	0.8298
0.511	8.99	832	0.5202	0.8845
0.5025	10.0	925	0.5260	0.8784
0.508	10.99	1017	0.5307	0.8632
0.4652	12.0	1110	0.6060	0.8480
0.4432	12.99	1202	0.5051	0.8845
0.3373	14.0	1295	0.8695	0.8845
0.3968	14.99	1387	0.6805	0.8571
0.4268	16.0	1480	0.6541	0.8815
0.3029	16.99	1572	0.5710	0.8906
0.3801	18.0	1665	0.6499	0.8571
0.3545	18.99	1757	0.6727	0.8419
0.3526	20.0	1850	0.6542	0.8571
0.3458	20.99	1942	0.6625	0.8997
0.3078	22.0	2035	0.6551	0.8784
0.3677	22.99	2127	0.5953	0.8815
0.3386	24.0	2220	0.6549	0.8693
0.213	24.99	2312	0.5846	0.8997
0.3778	26.0	2405	0.6746	0.8602
0.3079	26.99	2497	0.6594	0.8997
0.2943	28.0	2590	0.6246	0.8815
0.2782	28.99	2682	0.6550	0.8906
0.2931	29.84	2760	0.5844	0.8845

Framework versions

• Transformers 4.30.2
• Pytorch 2.0.1+cu117
• Datasets 2.13.0
• Tokenizers 0.13.3