segformer-b0-finetuned-segments-dots

This model is a fine-tuned version of nvidia/mit-b0 on the rohan8020/test dataset. It achieves the following results on the evaluation set:

• Loss: 0.1328
• Mean Iou: 0.3201
• Mean Accuracy: 0.6402
• Overall Accuracy: 0.6402
• Accuracy Unlabeled: nan
• Accuracy Dots: 0.6402
• Iou Unlabeled: 0.0
• Iou Dots: 0.6402

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 Dots	Iou Unlabeled	Iou Dots
0.5642	4.0	20	0.6209	0.4838	0.9677	0.9677	nan	0.9677	0.0	0.9677
0.4154	8.0	40	0.4119	0.2969	0.5939	0.5939	nan	0.5939	0.0	0.5939
0.3246	12.0	60	0.2900	0.3123	0.6246	0.6246	nan	0.6246	0.0	0.6246
0.2898	16.0	80	0.3168	0.4260	0.8520	0.8520	nan	0.8520	0.0	0.8520
0.2419	20.0	100	0.2201	0.3446	0.6892	0.6892	nan	0.6892	0.0	0.6892
0.2042	24.0	120	0.2199	0.3213	0.6426	0.6426	nan	0.6426	0.0	0.6426
0.1662	28.0	140	0.1797	0.3002	0.6005	0.6005	nan	0.6005	0.0	0.6005
0.1757	32.0	160	0.1611	0.2919	0.5839	0.5839	nan	0.5839	0.0	0.5839
0.1473	36.0	180	0.1477	0.3219	0.6439	0.6439	nan	0.6439	0.0	0.6439
0.1645	40.0	200	0.1448	0.3267	0.6534	0.6534	nan	0.6534	0.0	0.6534
0.1576	44.0	220	0.1389	0.3377	0.6754	0.6754	nan	0.6754	0.0	0.6754
0.1381	48.0	240	0.1328	0.3201	0.6402	0.6402	nan	0.6402	0.0	0.6402

Framework versions

• Transformers 4.37.0
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0