convnext-nano-20ep

This model is a fine-tuned version of facebook/convnextv2-nano-22k-384 on the vuongnhathien/30VNFoods dataset. It achieves the following results on the evaluation set:

• Loss: 0.4812
• Accuracy: 0.8702

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: 0.0003
• train_batch_size: 64
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.5831	1.0	275	0.5660	0.8278
0.3159	2.0	550	0.5093	0.8529
0.1892	3.0	825	0.4719	0.8779
0.1111	4.0	1100	0.5067	0.8755
0.0886	5.0	1375	0.5278	0.8708
0.0697	6.0	1650	0.6000	0.8628
0.0396	7.0	1925	0.6158	0.8736
0.0386	8.0	2200	0.6448	0.8684
0.0323	9.0	2475	0.5637	0.8915
0.0157	10.0	2750	0.5845	0.8958
0.0067	11.0	3025	0.5574	0.9018
0.005	12.0	3300	0.5378	0.9034
0.0031	13.0	3575	0.5526	0.9014
0.0023	14.0	3850	0.5419	0.9093
0.0026	15.0	4125	0.5323	0.9113
0.0024	16.0	4400	0.5298	0.9117
0.0019	17.0	4675	0.5323	0.9121
0.002	18.0	4950	0.5315	0.9125
0.0012	19.0	5225	0.5314	0.9121
0.0019	20.0	5500	0.5315	0.9117

Framework versions

• Transformers 4.39.3
• Pytorch 2.1.2
• Datasets 2.18.0
• Tokenizers 0.15.2