segformer-b0-finetuned-sudoku

This model is a fine-tuned version of nvidia/mit-b0 on the mrkprc1/SudokuBoundaries2 dataset. It achieves the following results on the evaluation set:

• Loss: 1.5465
• Mean Iou: 0.2407
• Mean Accuracy: 0.5
• Overall Accuracy: 0.4814
• Accuracy Unlabelled: 1.0
• Accuracy Sudoku-boundary: 0.0
• Iou Unlabelled: 0.4814
• Iou Sudoku-boundary: 0.0

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: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Unlabelled	Accuracy Sudoku-boundary	Iou Unlabelled	Iou Sudoku-boundary
0.6257	2.5	20	0.7024	0.2992	0.4856	0.4769	0.7186	0.2525	0.3981	0.2002
0.6194	5.0	40	0.7513	0.2593	0.4960	0.4797	0.9332	0.0588	0.4633	0.0553
0.6134	7.5	60	0.8649	0.2428	0.4993	0.4809	0.9921	0.0065	0.4792	0.0065
0.4962	10.0	80	0.9245	0.2434	0.5006	0.4822	0.9949	0.0063	0.4805	0.0063
0.5552	12.5	100	0.8606	0.2442	0.5009	0.4826	0.9939	0.0080	0.4804	0.0079
0.6282	15.0	120	1.1507	0.2407	0.5000	0.4814	1.0000	0.0000	0.4813	0.0000
0.4042	17.5	140	1.0916	0.2408	0.4997	0.4811	0.9988	0.0007	0.4810	0.0007
0.8174	20.0	160	0.9731	0.2424	0.4991	0.4807	0.9926	0.0056	0.4792	0.0055
0.5353	22.5	180	0.9754	0.2409	0.4991	0.4805	0.9964	0.0017	0.4801	0.0017
0.4792	25.0	200	1.6835	0.2407	0.5	0.4814	1.0	0.0	0.4814	0.0
0.4244	27.5	220	1.5039	0.2407	0.5	0.4814	1.0	0.0	0.4814	0.0
0.376	30.0	240	2.2746	0.2407	0.5	0.4814	1.0	0.0	0.4814	0.0
0.4129	32.5	260	2.0116	0.2407	0.5	0.4814	1.0	0.0	0.4814	0.0
0.4717	35.0	280	1.8957	0.2407	0.5	0.4814	1.0	0.0	0.4814	0.0
0.4229	37.5	300	1.7574	0.2407	0.5	0.4814	1.0	0.0	0.4814	0.0
0.5708	40.0	320	2.0764	0.2407	0.5	0.4814	1.0	0.0	0.4814	0.0
0.5826	42.5	340	1.6177	0.2407	0.5	0.4814	1.0	0.0	0.4814	0.0
0.3765	45.0	360	1.8119	0.2407	0.5	0.4814	1.0	0.0	0.4814	0.0
0.3704	47.5	380	1.6863	0.2407	0.5	0.4814	1.0	0.0	0.4814	0.0
1.3265	50.0	400	1.5465	0.2407	0.5	0.4814	1.0	0.0	0.4814	0.0

Framework versions

• Transformers 4.37.1
• Pytorch 2.1.2
• Datasets 2.16.1
• Tokenizers 0.15.1