segformer-b5-finetuned-ade20k-morphpadver1-hgo-coord-_40epochs_distortion_30_30_global_norm

This model is a fine-tuned version of NICOPOI-9/segformer-b5-finetuned-ade20k-morphpadver1-hgo-coord_40epochs_distortion_global_norm on the NICOPOI-9/Morphpad_HGO_1000_30_30_coord_global_norm dataset. It achieves the following results on the evaluation set:

• Loss: 0.3022
• Mean Iou: 0.8265
• Mean Accuracy: 0.9052
• Overall Accuracy: 0.9052
• Accuracy 0-0: 0.9412
• Accuracy 0-90: 0.8587
• Accuracy 90-0: 0.8790
• Accuracy 90-90: 0.9419
• Iou 0-0: 0.8523
• Iou 0-90: 0.7989
• Iou 90-0: 0.8052
• Iou 90-90: 0.8497

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: 6e-05
• train_batch_size: 1
• eval_batch_size: 1
• 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: linear
• num_epochs: 40

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy 0-0	Accuracy 0-90	Accuracy 90-0	Accuracy 90-90	Iou 0-0	Iou 0-90	Iou 90-0	Iou 90-90
1.3648	4.2105	4000	0.6460	0.6389	0.7795	0.7795	0.8410	0.7043	0.7272	0.8455	0.6851	0.5820	0.5850	0.7036
1.3689	8.4211	8000	0.5698	0.7166	0.8342	0.8341	0.8712	0.7811	0.8184	0.8660	0.7602	0.6620	0.6760	0.7683
0.2883	12.6316	12000	0.3956	0.7572	0.8614	0.8614	0.8961	0.8453	0.8126	0.8916	0.7990	0.7202	0.7123	0.7972
0.2494	16.8421	16000	0.3556	0.7756	0.8736	0.8737	0.9125	0.8340	0.8208	0.9271	0.8166	0.7384	0.7369	0.8104
1.3766	21.0526	20000	0.3541	0.7927	0.8845	0.8844	0.9252	0.8372	0.8541	0.9214	0.8232	0.7589	0.7623	0.8264
0.1549	25.2632	24000	0.3087	0.8151	0.8982	0.8981	0.9363	0.8528	0.8764	0.9273	0.8455	0.7821	0.7863	0.8465
1.3569	29.4737	28000	0.3299	0.8130	0.8971	0.8971	0.9416	0.8450	0.8653	0.9366	0.8415	0.7811	0.7883	0.8412
1.3663	33.6842	32000	0.2746	0.8337	0.9094	0.9094	0.9383	0.8793	0.8755	0.9444	0.8607	0.8076	0.8073	0.8593
0.6055	37.8947	36000	0.3022	0.8265	0.9052	0.9052	0.9412	0.8587	0.8790	0.9419	0.8523	0.7989	0.8052	0.8497

Framework versions

• Transformers 4.48.3
• Pytorch 2.1.0
• Datasets 3.2.0
• Tokenizers 0.21.0