segformer-b0-finetuned-segments-sidewalk-123

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

• eval_loss: 0.5603
• eval_mean_iou: 0.3194
• eval_mean_accuracy: 0.3797
• eval_overall_accuracy: 0.8649
• eval_accuracy_unlabeled: nan
• eval_accuracy_flat-road: 0.8652
• eval_accuracy_flat-sidewalk: 0.9642
• eval_accuracy_flat-crosswalk: 0.6083
• eval_accuracy_flat-cyclinglane: 0.7173
• eval_accuracy_flat-parkingdriveway: 0.5189
• eval_accuracy_flat-railtrack: 0.0
• eval_accuracy_flat-curb: 0.6089
• eval_accuracy_human-person: 0.8421
• eval_accuracy_human-rider: 0.0
• eval_accuracy_vehicle-car: 0.9347
• eval_accuracy_vehicle-truck: 0.0
• eval_accuracy_vehicle-bus: 0.0
• eval_accuracy_vehicle-tramtrain: 0.0
• eval_accuracy_vehicle-motorcycle: 0.0
• eval_accuracy_vehicle-bicycle: 0.7877
• eval_accuracy_vehicle-caravan: 0.0
• eval_accuracy_vehicle-cartrailer: 0.0
• eval_accuracy_construction-building: 0.9288
• eval_accuracy_construction-door: 0.0851
• eval_accuracy_construction-wall: 0.4696
• eval_accuracy_construction-fenceguardrail: 0.4756
• eval_accuracy_construction-bridge: 0.0
• eval_accuracy_construction-tunnel: nan
• eval_accuracy_construction-stairs: 0.0047
• eval_accuracy_object-pole: 0.3599
• eval_accuracy_object-trafficsign: 0.0028
• eval_accuracy_object-trafficlight: 0.0
• eval_accuracy_nature-vegetation: 0.9419
• eval_accuracy_nature-terrain: 0.8921
• eval_accuracy_sky: 0.9773
• eval_accuracy_void-ground: 0.0003
• eval_accuracy_void-dynamic: 0.1852
• eval_accuracy_void-static: 0.3603
• eval_accuracy_void-unclear: 0.0
• eval_iou_unlabeled: nan
• eval_iou_flat-road: 0.7774
• eval_iou_flat-sidewalk: 0.8659
• eval_iou_flat-crosswalk: 0.4889
• eval_iou_flat-cyclinglane: 0.6488
• eval_iou_flat-parkingdriveway: 0.4072
• eval_iou_flat-railtrack: 0.0
• eval_iou_flat-curb: 0.4944
• eval_iou_human-person: 0.6064
• eval_iou_human-rider: 0.0
• eval_iou_vehicle-car: 0.8283
• eval_iou_vehicle-truck: 0.0
• eval_iou_vehicle-bus: 0.0
• eval_iou_vehicle-tramtrain: 0.0
• eval_iou_vehicle-motorcycle: 0.0
• eval_iou_vehicle-bicycle: 0.5105
• eval_iou_vehicle-caravan: 0.0
• eval_iou_vehicle-cartrailer: 0.0
• eval_iou_construction-building: 0.7458
• eval_iou_construction-door: 0.0760
• eval_iou_construction-wall: 0.4006
• eval_iou_construction-fenceguardrail: 0.3288
• eval_iou_construction-bridge: 0.0
• eval_iou_construction-tunnel: nan
• eval_iou_construction-stairs: 0.0047
• eval_iou_object-pole: 0.2789
• eval_iou_object-trafficsign: 0.0027
• eval_iou_object-trafficlight: 0.0
• eval_iou_nature-vegetation: 0.8781
• eval_iou_nature-terrain: 0.8170
• eval_iou_sky: 0.9417
• eval_iou_void-ground: 0.0002
• eval_iou_void-dynamic: 0.1706
• eval_iou_void-static: 0.2686
• eval_iou_void-unclear: 0.0
• eval_runtime: 246.8492
• eval_samples_per_second: 0.81
• eval_steps_per_second: 0.405
• epoch: 15.45
• step: 6180

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

Framework versions

• Transformers 4.35.0
• Pytorch 2.2.0+cpu
• Datasets 2.14.6
• Tokenizers 0.14.1