README.md

metadata

license: other
tags:
  - generated_from_keras_callback
model-index:
  - name: AhamadShaik/SegFormer_RESIZE_LM
    results: []

AhamadShaik/SegFormer_RESIZE_LM

This model is a fine-tuned version of nvidia/mit-b0 on an unknown dataset. It achieves the following results on the evaluation set:

• Train Loss: 0.0679
• Train Dice Coef: 0.7980
• Train Iou: 0.6817
• Validation Loss: 0.0483
• Validation Dice Coef: 0.8809
• Validation Iou: 0.7881
• Train Lr: 5e-06
• Epoch: 13

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:

• optimizer: {'name': 'Adam', 'learning_rate': 5e-06, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False}
• training_precision: float32

Training results

Train Loss	Train Dice Coef	Train Iou	Validation Loss	Validation Dice Coef	Validation Iou	Train Lr	Epoch
0.3496	0.3697	0.2435	0.2697	0.1141	0.0635	1e-04	0
0.1591	0.5600	0.4126	0.1768	0.3601	0.2345	1e-04	1
0.1295	0.6470	0.5014	0.1637	0.4628	0.3163	1e-04	2
0.1109	0.6903	0.5511	0.1319	0.5634	0.4072	1e-04	3
0.1018	0.7226	0.5858	0.0932	0.7480	0.6051	1e-04	4
0.0930	0.7373	0.6042	0.1618	0.5048	0.3614	1e-04	5
0.0878	0.7534	0.6255	0.1023	0.7076	0.5637	1e-04	6
0.0842	0.7585	0.6310	0.0878	0.7726	0.6384	1e-04	7
0.0798	0.7733	0.6475	0.0966	0.7434	0.5996	1e-04	8
0.0765	0.7716	0.6487	0.1073	0.7157	0.5657	1e-04	9
0.0701	0.7974	0.6794	0.1049	0.7190	0.5811	1e-04	10
0.0675	0.8020	0.6854	0.1319	0.6935	0.5427	1e-04	11
0.0662	0.8108	0.6957	0.1593	0.6269	0.4826	1e-04	12
0.0679	0.7980	0.6817	0.0483	0.8809	0.7881	5e-06	13

Framework versions

• Transformers 4.27.4
• TensorFlow 2.10.1
• Datasets 2.11.0
• Tokenizers 0.13.3