parking-utcustom-train-SF-RGB-b5_3

This model is a fine-tuned version of nvidia/mit-b5 on the sam1120/parking-utcustom-train dataset. It achieves the following results on the evaluation set:

• Loss: 0.1065
• Mean Iou: 1.0
• Mean Accuracy: 1.0
• Overall Accuracy: 1.0
• Accuracy Unlabeled: nan
• Accuracy Parking: nan
• Accuracy Unparking: 1.0
• Iou Unlabeled: nan
• Iou Parking: nan
• Iou Unparking: 1.0

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: 2.5e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 120

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Unlabeled	Accuracy Parking	Accuracy Unparking	Iou Unlabeled	Iou Parking	Iou Unparking
0.8229	20.0	20	0.7131	0.4707	0.9415	0.9415	nan	nan	0.9415	0.0	nan	0.9415
0.4809	40.0	40	0.3982	0.4905	0.9809	0.9809	nan	nan	0.9809	0.0	nan	0.9809
0.2948	60.0	60	0.2903	0.3300	0.9899	0.9899	nan	nan	0.9899	0.0	0.0	0.9899
0.227	80.0	80	0.1803	1.0	1.0	1.0	nan	nan	1.0	nan	nan	1.0
0.1736	100.0	100	0.1304	1.0	1.0	1.0	nan	nan	1.0	nan	nan	1.0
0.1568	120.0	120	0.1065	1.0	1.0	1.0	nan	nan	1.0	nan	nan	1.0

Framework versions

• Transformers 4.30.2
• Pytorch 2.0.1+cu117
• Datasets 2.13.1
• Tokenizers 0.13.3