SegFormer_b2_10

This model is a fine-tuned version of nvidia/segformer-b2-finetuned-cityscapes-1024-1024 on an unknown dataset. It achieves the following results on the evaluation set:

• epoch: 14.5161
• eval_accuracy_bicycle: 0.8914
• eval_accuracy_building: 0.9612
• eval_accuracy_bus: 0.9483
• eval_accuracy_car: 0.9763
• eval_accuracy_fence: 0.7181
• eval_accuracy_motorcycle: 0.7986
• eval_accuracy_person: 0.9057
• eval_accuracy_pole: 0.7198
• eval_accuracy_rider: 0.7552
• eval_accuracy_road: 0.9902
• eval_accuracy_sidewalk: 0.9345
• eval_accuracy_sky: 0.9831
• eval_accuracy_terrain: 0.7525
• eval_accuracy_traffic light: 0.8652
• eval_accuracy_traffic sign: 0.8838
• eval_accuracy_train: 0.8680
• eval_accuracy_truck: 0.8765
• eval_accuracy_vegetation: 0.9637
• eval_accuracy_wall: 0.7237
• eval_iou_bicycle: 0.7541
• eval_iou_building: 0.9244
• eval_iou_bus: 0.8603
• eval_iou_car: 0.9482
• eval_iou_fence: 0.6075
• eval_iou_motorcycle: 0.6289
• eval_iou_person: 0.7921
• eval_iou_pole: 0.5893
• eval_iou_rider: 0.5955
• eval_iou_road: 0.9835
• eval_iou_sidewalk: 0.8649
• eval_iou_sky: 0.9465
• eval_iou_terrain: 0.6534
• eval_iou_traffic light: 0.6718
• eval_iou_traffic sign: 0.7801
• eval_iou_train: 0.8124
• eval_iou_truck: 0.8174
• eval_iou_vegetation: 0.9245
• eval_iou_wall: 0.6499
• eval_loss: 0.8030
• eval_mean_accuracy: 0.8693
• eval_mean_iou: 0.7792
• eval_overall_accuracy: 0.9609
• eval_runtime: 202.8122
• eval_samples_per_second: 2.465
• eval_steps_per_second: 0.616
• step: 2700

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• 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
• lr_scheduler_warmup_steps: 1000
• num_epochs: 20
• mixed_precision_training: Native AMP

Framework versions

• Transformers 4.48.1
• Pytorch 2.1.2+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0