segformer-b0-finetuned-morphpadver1-hgo-30-coord-v3_2

This model is a fine-tuned version of nvidia/mit-b3 on the NICOPOI-9/morphpad_coord_hgo_30_30_512_4class dataset. It achieves the following results on the evaluation set:

• Loss: 0.7675
• Mean Iou: 0.4538
• Mean Accuracy: 0.6212
• Overall Accuracy: 0.6213
• Accuracy 0-0: 0.5825
• Accuracy 0-90: 0.5481
• Accuracy 90-0: 0.7589
• Accuracy 90-90: 0.5953
• Iou 0-0: 0.4693
• Iou 0-90: 0.4525
• Iou 90-0: 0.4219
• Iou 90-90: 0.4714

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: 1
• eval_batch_size: 1
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy 0-0	Accuracy 0-90	Accuracy 90-0	Accuracy 90-90	Iou 0-0	Iou 0-90	Iou 90-0	Iou 90-90
1.1766	4.2105	4000	1.1683	0.2583	0.4140	0.4143	0.2931	0.3809	0.6404	0.3417	0.2408	0.2494	0.2769	0.2663
1.0168	8.4211	8000	1.0180	0.3414	0.5100	0.5089	0.7153	0.4170	0.4442	0.4636	0.3414	0.3305	0.3283	0.3653
0.8472	12.6316	12000	0.8536	0.4077	0.5792	0.5793	0.6031	0.6125	0.4758	0.6254	0.4108	0.3947	0.4040	0.4211
1.0326	16.8421	16000	0.7675	0.4538	0.6212	0.6213	0.5825	0.5481	0.7589	0.5953	0.4693	0.4525	0.4219	0.4714

Framework versions

• Transformers 4.48.3
• Pytorch 2.1.0
• Datasets 3.2.0
• Tokenizers 0.21.0