segformerSAAD

This model is a fine-tuned version of nvidia/mit-b0 on the saad7489/SixGUN dataset. It achieves the following results on the evaluation set:

• Loss: 0.7611
• Mean Iou: 0.5823
• Mean Accuracy: 0.8994
• Overall Accuracy: 0.9474
• Accuracy Bkg: 0.9505
• Accuracy Knife: 0.8767
• Accuracy Gun: 0.8711
• Iou Bkg: 0.9471
• Iou Knife: 0.4452
• Iou Gun: 0.3544

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: 20
• eval_batch_size: 20
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Bkg	Accuracy Knife	Accuracy Gun	Iou Bkg	Iou Knife	Iou Gun
0.9115	10.0	20	1.0252	0.5076	0.9065	0.9181	0.9188	0.8625	0.9382	0.9176	0.3218	0.2833
0.766	20.0	40	0.8278	0.5811	0.8914	0.9486	0.9523	0.8691	0.8527	0.9485	0.4288	0.3661
0.7862	30.0	60	0.7611	0.5823	0.8994	0.9474	0.9505	0.8767	0.8711	0.9471	0.4452	0.3544

Framework versions

• Transformers 4.42.4
• Pytorch 2.3.1+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1