resnet-101-finetuned_resnet101-sgd-optimizer20-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: 0.3318
• Accuracy: 0.8848

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: 0.1
• 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
1.1302	0.99	65	1.0040	0.6724
1.1708	1.99	130	1.4856	0.5495
1.141	2.99	195	1.1486	0.6352
1.0119	3.99	260	0.8829	0.7314
0.8091	4.99	325	0.8301	0.7419
0.7878	5.99	390	0.8121	0.7333
0.6827	6.99	455	0.6047	0.7990
0.5525	7.99	520	0.6028	0.8048
0.5787	8.99	585	0.5183	0.8352
0.4797	9.99	650	0.4737	0.8543
0.4224	10.99	715	0.4943	0.8305
0.4389	11.99	780	0.4162	0.8629
0.4142	12.99	845	0.4000	0.8629
0.3144	13.99	910	0.3833	0.8695
0.2915	14.99	975	0.3688	0.8733
0.3302	15.99	1040	0.3643	0.8810
0.2954	16.99	1105	0.3446	0.8867
0.2186	17.99	1170	0.3571	0.8905
0.1812	18.99	1235	0.3334	0.8886
0.1911	19.99	1300	0.3318	0.8848

Framework versions

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