MobileNet-V2-food

This model is a fine-tuned version of google/mobilenet_v2_1.0_224 on the ItsNotRohit/Food121-224 dataset. It achieves the following results on the evaluation set:

• Loss: 1.6890
• Accuracy: 0.5793
• Recall: 0.5793
• Precision: 0.6006
• F1: 0.5769

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: 16
• eval_batch_size: 128
• seed: 20329
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• training_steps: 20000

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Recall	Precision	F1
2.9653	0.33	2000	2.7802	0.3438	0.3438	0.3932	0.3105
2.3854	0.66	4000	2.3105	0.4440	0.4440	0.4979	0.4336
2.1576	0.99	6000	2.0508	0.4958	0.4958	0.5263	0.4837
1.9767	1.32	8000	1.9860	0.5086	0.5086	0.5504	0.4956
1.9215	1.65	10000	1.8312	0.5462	0.5462	0.5815	0.5390
1.782	1.98	12000	1.8554	0.5441	0.5441	0.5864	0.5431
1.7755	2.31	14000	1.9241	0.5308	0.5308	0.5841	0.5272
1.7006	2.64	16000	1.8625	0.5451	0.5451	0.6004	0.5466
1.7289	2.98	18000	1.8560	0.5432	0.5432	0.5940	0.5395
1.7296	3.31	20000	1.6890	0.5793	0.5793	0.6006	0.5769

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0