resnet-152_rice-leaf-disease-augmented-v3_fft

This model is a fine-tuned version of microsoft/resnet-152 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.7666
• Accuracy: 0.8567

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: 64
• eval_batch_size: 64
• 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
2.051	1.0	128	1.9776	0.4560
1.8042	2.0	256	1.4858	0.5993
1.129	3.0	384	0.8132	0.7459
0.6016	4.0	512	0.6129	0.8111
0.3757	5.0	640	0.5488	0.8274
0.2893	6.0	768	0.5463	0.8469
0.2625	7.0	896	0.5314	0.8469
0.1559	8.0	1024	0.4619	0.8664
0.0772	9.0	1152	0.5055	0.8567
0.0471	10.0	1280	0.5659	0.8567
0.0359	11.0	1408	0.5393	0.8534
0.0346	12.0	1536	0.6242	0.8502
0.0205	13.0	1664	0.6405	0.8632
0.0138	14.0	1792	0.6308	0.8632
0.0102	15.0	1920	0.6098	0.8730
0.0099	16.0	2048	0.7251	0.8599
0.0072	17.0	2176	0.7383	0.8534
0.0055	18.0	2304	0.8013	0.8567
0.0033	19.0	2432	0.7485	0.8599
0.0038	20.0	2560	0.7666	0.8567

Framework versions

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