mobilevit-finetuned-food101

This model is a fine-tuned version of apple/mobilevitv2-1.0-imagenet1k-256 on the food101 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4191
• Accuracy: 0.874

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• 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
1.9487	0.98	23	1.9476	0.151
1.9273	2.0	47	1.9070	0.24
1.8561	2.98	70	1.8401	0.448
1.7788	4.0	94	1.7301	0.612
1.6586	4.98	117	1.5863	0.676
1.4603	6.0	141	1.4199	0.72
1.3027	6.98	164	1.2215	0.734
1.1717	8.0	188	1.0581	0.759
0.9601	8.98	211	0.9013	0.769
0.8482	10.0	235	0.7866	0.791
0.7276	10.98	258	0.7112	0.803
0.6449	12.0	282	0.6132	0.835
0.6279	12.98	305	0.6069	0.83
0.5982	14.0	329	0.5637	0.832
0.5766	14.98	352	0.5149	0.857
0.5345	16.0	376	0.5392	0.837
0.494	16.98	399	0.5017	0.848
0.4953	18.0	423	0.5002	0.846
0.5118	18.98	446	0.4782	0.856
0.4708	20.0	470	0.4898	0.858
0.4774	20.98	493	0.4769	0.851
0.4848	22.0	517	0.4665	0.841
0.4533	22.98	540	0.4890	0.837
0.4449	24.0	564	0.4558	0.857
0.4205	24.98	587	0.4767	0.857
0.4417	26.0	611	0.4476	0.853
0.4333	26.98	634	0.4853	0.834
0.4545	28.0	658	0.4573	0.847
0.4489	28.98	681	0.4659	0.845
0.4172	29.36	690	0.4191	0.874

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0