resnet-101-finetuned_resnet101-adam-optimizer5e-4-autotags

This model is a fine-tuned version of microsoft/resnet-101 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.2477
• Accuracy: 0.9267

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: 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
2.6033	0.99	65	2.5693	0.1381
1.5517	1.99	130	1.1376	0.6733
0.9423	2.99	195	0.6290	0.7895
0.6334	3.99	260	0.4372	0.86
0.4735	4.99	325	0.4719	0.8429
0.4573	5.99	390	0.3909	0.8590
0.3236	6.99	455	0.3507	0.8752
0.2511	7.99	520	0.2931	0.9019
0.2073	8.99	585	0.2757	0.9133
0.2174	9.99	650	0.2706	0.9114
0.1558	10.99	715	0.2654	0.9114
0.2017	11.99	780	0.2820	0.9114
0.134	12.99	845	0.2431	0.9238
0.0943	13.99	910	0.2606	0.9105
0.1396	14.99	975	0.2514	0.9229
0.1374	15.99	1040	0.2349	0.9305
0.0953	16.99	1105	0.2502	0.9210
0.0742	17.99	1170	0.2515	0.9210
0.0708	18.99	1235	0.2437	0.9257
0.0619	19.99	1300	0.2477	0.9267

Framework versions

• Transformers 4.25.1
• Pytorch 1.13.1+cu117
• Datasets 2.11.0
• Tokenizers 0.13.2