README.md

metadata

license: other
tags:
  - vision
  - image-segmentation
  - generated_from_trainer
base_model: nvidia/mit-b1
model-index:
  - name: segformer-b1-finetuned-sudoku
    results: []

segformer-b1-finetuned-sudoku

This model is a fine-tuned version of nvidia/mit-b1 on the mrkprc1/SudokuBoundaries2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.7703
• Mean Iou: 0.0967
• Mean Accuracy: 0.1934
• Overall Accuracy: 0.1934
• Accuracy Unlabelled: nan
• Accuracy Sudoku-boundary: 0.1934
• Iou Unlabelled: 0.0
• Iou Sudoku-boundary: 0.1934

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
0.6531	3.33	20	0.7016	0.1433	0.2867	0.2867	nan	0.2867	0.0	0.2867
0.7654	6.67	40	0.7142	0.3064	0.6129	0.6129	nan	0.6129	0.0	0.6129
0.4761	10.0	60	1.0391	0.0002	0.0005	0.0005	nan	0.0005	0.0	0.0005
0.7746	13.33	80	1.7648	0.0	0.0	0.0	nan	0.0	0.0	0.0
0.5488	16.67	100	1.2288	0.0	0.0	0.0	nan	0.0	0.0	0.0
0.6242	20.0	120	1.5012	0.0	0.0	0.0	nan	0.0	0.0	0.0
0.5423	23.33	140	0.9650	0.0029	0.0059	0.0059	nan	0.0059	0.0	0.0059
0.521	26.67	160	0.8594	0.0197	0.0393	0.0393	nan	0.0393	0.0	0.0393
0.5655	30.0	180	0.7950	0.0527	0.1055	0.1055	nan	0.1055	0.0	0.1055
0.4229	33.33	200	0.7910	0.0982	0.1964	0.1964	nan	0.1964	0.0	0.1964
0.288	36.67	220	0.7591	0.1358	0.2715	0.2715	nan	0.2715	0.0	0.2715
0.2002	40.0	240	0.7395	0.2414	0.4828	0.4828	nan	0.4828	0.0	0.4828
0.6014	43.33	260	0.7405	0.2644	0.5289	0.5289	nan	0.5289	0.0	0.5289
0.4336	46.67	280	0.7423	0.1751	0.3502	0.3502	nan	0.3502	0.0	0.3502
0.565	50.0	300	0.7703	0.0967	0.1934	0.1934	nan	0.1934	0.0	0.1934

Framework versions

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