convnextv2-base-1k-224-finetuned-cassava-leaf-disease-randomflip

This model is a fine-tuned version of facebook/convnextv2-base-1k-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.3704
• Accuracy: 0.8766

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: 400
• eval_batch_size: 400
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 1600
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 16

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.5605	0.98	12	1.2754	0.6150
1.2015	1.96	24	0.9009	0.6290
0.9048	2.94	36	0.6987	0.7701
0.7362	4.0	49	0.5497	0.8206
0.5294	4.98	61	0.4712	0.8542
0.4777	5.96	73	0.4451	0.8547
0.458	6.94	85	0.4197	0.8579
0.4208	8.0	98	0.4084	0.8682
0.4042	8.98	110	0.3930	0.8692
0.4071	9.96	122	0.3879	0.8743
0.3868	10.94	134	0.3923	0.8715
0.3849	12.0	147	0.3763	0.8748
0.3744	12.98	159	0.3732	0.8776
0.3739	13.96	171	0.3708	0.8748
0.361	14.94	183	0.3693	0.8818
0.3725	15.67	192	0.3704	0.8766

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.1
• Datasets 2.18.0
• Tokenizers 0.15.1