segformer-finetuned

This model is a fine-tuned version of nvidia/mit-b0 on an unknown dataset.

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: 3e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• 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: cosine
• lr_scheduler_warmup_steps: 50
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Dice	Mean Precision	Mean Recall	Pixel Acc	Class 0 Iou	Class 0 Dice	Class 0 Precision	Class 0 Recall	Class 1 Iou	Class 1 Dice	Class 1 Precision	Class 1 Recall
No log	4.1739	50	0.1130	0.7371	0.8240	0.8335	0.8151	0.9885	0.9884	0.9942	0.9937	0.9946	0.4858	0.6539	0.6733	0.6355

Framework versions

• Transformers 4.57.0
• Pytorch 2.6.0
• Datasets 2.2.1
• Tokenizers 0.22.1