metadata

library_name: transformers
license: other
base_model: nvidia/mit-b0
tags:
  - vision
  - image-segmentation
  - generated_from_trainer
model-index:
  - name: segformer-b0-finetuned-test
    results: []

segformer-b0-finetuned-test

This model is a fine-tuned version of nvidia/mit-b0 on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.2090
Mean Iou: 0.5696
Mean Accuracy: 0.6654
Overall Accuracy: 0.9106
Accuracy Structure (dimensional): nan
Accuracy Impervious (planiform): 0.9401
Accuracy Fences: 0.0
Accuracy Water storage/tank: nan
Accuracy Pool < 100 sqft: nan
Accuracy Pool > 100 sqft: 0.9530
Accuracy Irrigated planiform: 0.8838
Accuracy Irrigated dimensional low: 0.8370
Accuracy Irrigated dimensional high: 0.9432
Accuracy Irrigated bare: 0.4234
Accuracy Irrigable planiform: 0.8112
Accuracy Irrigable dimensional low: 0.5718
Accuracy Irrigable dimensional high: 0.9410
Accuracy Irrigable bare: 0.7245
Accuracy Native planiform: nan
Accuracy Native dimensional low: 0.0
Accuracy Native dimensional high: 0.0
Accuracy Native bare: 0.9472
Accuracy Udl: nan
Accuracy Open water: 0.9967
Accuracy Artificial turf: 0.6743
Iou Structure (dimensional): 0.0
Iou Impervious (planiform): 0.8873
Iou Fences: 0.0
Iou Water storage/tank: nan
Iou Pool < 100 sqft: nan
Iou Pool > 100 sqft: 0.8999
Iou Irrigated planiform: 0.7859
Iou Irrigated dimensional low: 0.7122
Iou Irrigated dimensional high: 0.8937
Iou Irrigated bare: 0.3648
Iou Irrigable planiform: 0.7323
Iou Irrigable dimensional low: 0.4772
Iou Irrigable dimensional high: 0.8844
Iou Irrigable bare: 0.6417
Iou Native planiform: nan
Iou Native dimensional low: 0.0
Iou Native dimensional high: 0.0
Iou Native bare: 0.8356
Iou Udl: nan
Iou Open water: 0.9389
Iou Artificial turf: 0.6287

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: 16
eval_batch_size: 16
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Structure (dimensional)	Accuracy Impervious (planiform)	Accuracy Fences	Accuracy Water storage/tank	Accuracy Pool < 100 sqft	Accuracy Pool > 100 sqft	Accuracy Irrigated planiform	Accuracy Irrigated dimensional low	Accuracy Irrigated dimensional high	Accuracy Irrigated bare	Accuracy Irrigable planiform	Accuracy Irrigable dimensional low	Accuracy Irrigable dimensional high	Accuracy Irrigable bare	Accuracy Native planiform	Accuracy Native dimensional low	Accuracy Native dimensional high	Accuracy Native bare	Accuracy Udl	Accuracy Open water	Accuracy Artificial turf	Iou Structure (dimensional)	Iou Impervious (planiform)	Iou Fences	Iou Water storage/tank	Iou Pool < 100 sqft	Iou Pool > 100 sqft	Iou Irrigated planiform	Iou Irrigated dimensional low	Iou Irrigated dimensional high	Iou Irrigated bare	Iou Irrigable planiform	Iou Irrigable dimensional low	Iou Irrigable dimensional high	Iou Irrigable bare	Iou Native planiform	Iou Native dimensional low	Iou Native dimensional high	Iou Native bare	Iou Udl	Iou Open water	Iou Artificial turf
0.2526	15.3846	200	0.2090	0.5696	0.6654	0.9106	nan	0.9401	0.0	nan	nan	0.9530	0.8838	0.8370	0.9432	0.4234	0.8112	0.5718	0.9410	0.7245	nan	0.0	0.0	0.9472	nan	0.9967	0.6743	0.0	0.8873	0.0	nan	nan	0.8999	0.7859	0.7122	0.8937	0.3648	0.7323	0.4772	0.8844	0.6417	nan	0.0	0.0	0.8356	nan	0.9389	0.6287

Framework versions

Transformers 4.44.2
Pytorch 2.5.0+cu121
Datasets 3.1.0
Tokenizers 0.19.1