README.md

metadata

license: other
tags:
  - vision
  - image-segmentation
  - generated_from_trainer
model-index:
  - name: segformer-b0-finetuned-segments-toolwear
    results: []

segformer-b0-finetuned-segments-toolwear

This model is a fine-tuned version of nvidia/mit-b0 on the HorcruxNo13/new_wear dataset. It achieves the following results on the evaluation set:

• Loss: 0.0737
• Mean Iou: 0.3080
• Mean Accuracy: 0.6160
• Overall Accuracy: 0.6160
• Accuracy Unlabeled: nan
• Accuracy Wear: 0.6160
• Iou Unlabeled: 0.0
• Iou Wear: 0.6160

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 Unlabeled	Accuracy Wear	Iou Unlabeled	Iou Wear
0.4694	3.33	20	0.4857	0.3178	0.6356	0.6356	nan	0.6356	0.0	0.6356
0.3038	6.67	40	0.2805	0.3408	0.6816	0.6816	nan	0.6816	0.0	0.6816
0.2303	10.0	60	0.2080	0.3408	0.6816	0.6816	nan	0.6816	0.0	0.6816
0.1935	13.33	80	0.1870	0.3420	0.6841	0.6841	nan	0.6841	0.0	0.6841
0.1697	16.67	100	0.1507	0.3405	0.6810	0.6810	nan	0.6810	0.0	0.6810
0.1406	20.0	120	0.1377	0.3437	0.6874	0.6874	nan	0.6874	0.0	0.6874
0.1363	23.33	140	0.1156	0.3301	0.6601	0.6601	nan	0.6601	0.0	0.6601
0.117	26.67	160	0.1019	0.3376	0.6753	0.6753	nan	0.6753	0.0	0.6753
0.0972	30.0	180	0.0935	0.3264	0.6529	0.6529	nan	0.6529	0.0	0.6529
0.1076	33.33	200	0.0901	0.3292	0.6584	0.6584	nan	0.6584	0.0	0.6584
0.0868	36.67	220	0.0806	0.3218	0.6436	0.6436	nan	0.6436	0.0	0.6436
0.0866	40.0	240	0.0766	0.3183	0.6367	0.6367	nan	0.6367	0.0	0.6367
0.0757	43.33	260	0.0750	0.3082	0.6165	0.6165	nan	0.6165	0.0	0.6165
0.077	46.67	280	0.0750	0.3104	0.6207	0.6207	nan	0.6207	0.0	0.6207
0.0765	50.0	300	0.0737	0.3080	0.6160	0.6160	nan	0.6160	0.0	0.6160

Framework versions

• Transformers 4.28.0
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.13.3