Modelpublic

Omar280X/passport_segmentation_v9

sourceHugging Faceotherupdated 2y agoView on Hugging Face

0likes2downloads

Model Card

Omar280X/passportsegmentationv9

This model is a fine-tuned version of nvidia/segformer-b0-finetuned-ade-512-512 on an unknown dataset. It achieves the following results on the evaluation set:

—Train Loss: 0.0239
—Validation Loss: 0.0941
—Epoch: 19

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', 'learningrate': 6e-05, 'decay': 0.0, 'beta1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False}
—training_precision: float32

Training results

Train Loss	Validation Loss	Epoch
0.1594	0.1322	0
0.0783	0.1326	1
0.0620	0.1099	2
0.0531	0.1093	3
0.0484	0.0908	4
0.0440	0.1785	5
0.0403	0.1345	6
0.0379	0.0885	7
0.0357	0.1485	8
0.0334	0.1024	9
0.0308	0.1009	10
0.0311	0.1368	11
0.0321	0.1223	12
0.0290	0.0974	13
0.0272	0.1802	14
0.0276	0.1109	15
0.0260	0.1074	16
0.0279	0.1092	17
0.0252	0.0994	18
0.0239	0.0941	19

Framework versions

—Transformers 4.37.0
—TensorFlow 2.10.0
—Datasets 3.2.0
—Tokenizers 0.15.2