cvmil/convnext-base-224_augmented-v2_tl

convnext-base-224_rice-leaf-disease-augmented-v2_tl

This model is a fine-tuned version of facebook/convnext-base-224 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.9910
• Accuracy: 0.6935

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: 128
• eval_batch_size: 128
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine_with_restarts
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.0641	1.0	63	1.9738	0.3780
1.833	2.0	126	1.7100	0.5
1.5486	3.0	189	1.4962	0.5595
1.3602	4.0	252	1.3563	0.5982
1.2385	5.0	315	1.2619	0.625
1.1522	6.0	378	1.1995	0.6458
1.0888	7.0	441	1.1555	0.6607
1.0383	8.0	504	1.1159	0.6667
1.0003	9.0	567	1.0857	0.6726
0.971	10.0	630	1.0579	0.6845
0.9447	11.0	693	1.0446	0.6815
0.9265	12.0	756	1.0262	0.6905
0.9123	13.0	819	1.0158	0.6845
0.8985	14.0	882	1.0050	0.6935
0.8902	15.0	945	0.9993	0.6964
0.8846	16.0	1008	0.9950	0.6935
0.8803	17.0	1071	0.9926	0.6935
0.8774	18.0	1134	0.9916	0.6935
0.8766	19.0	1197	0.9911	0.6935
0.8772	20.0	1260	0.9910	0.6935

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0