segformer-b0-finetuned-segments-sidewalk-4

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:

• Loss: 2.5207
• Mean Iou: 0.1023
• Mean Accuracy: 0.1567
• Overall Accuracy: 0.6612
• Per Category Iou: [0.0, 0.37997208823402434, 0.7030895600821837, 0.0, 0.0020740824048893942, 0.0006611109803275343, 0.0, 0.0009644717061794479, 0.0, 0.0, 0.44780560238339745, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.4962679673706645, 0.0, 0.008267299447856608, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.6719286019431624, 0.1932540547332544, 0.6762198255750292, 0.0, 0.0, 0.0003312368464636427, 0.0]
• Per Category Accuracy: [nan, 0.7085417733756095, 0.8643251797889624, 0.0, 0.0020922282164545967, 0.0006691672739475508, nan, 0.0009725011389865425, 0.0, 0.0, 0.9224475476880146, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.7984415122785299, 0.0, 0.008394275137866055, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.9294223049507054, 0.2306496542338313, 0.7045666997791757, 0.0, 0.0, 0.0003315891206418271, 0.0]

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Per Category Iou	Per Category Accuracy
2.8255	1.0	25	3.0220	0.0892	0.1429	0.6352	[0.0, 0.3631053229188519, 0.6874502125236047, 0.0, 0.012635239862746197, 0.001133215250040838, 0.0, 0.00463024415429387, 2.6557099661207286e-05, 0.0, 0.3968535016422742, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.4820466790242289, 0.0, 0.00693999220077067, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.6134928158666486, 0.05160593984758798, 0.5016270369795023, 0.0, 0.0, 0.00023524914354608678, 0.0]	[nan, 0.6625398055826, 0.851744092156527, 0.0, 0.01307675614921835, 0.001170877257777663, nan, 0.004771009467501389, 2.6941417811356193e-05, 0.0, 0.9316713675735513, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.7310221003907382, 0.0, 0.0070371168820434, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.948375993368795, 0.056265031783493576, 0.5061367774453964, 0.0, 0.0, 0.00023723449281691698, 0.0]
2.5443	2.0	50	2.5207	0.1023	0.1567	0.6612	[0.0, 0.37997208823402434, 0.7030895600821837, 0.0, 0.0020740824048893942, 0.0006611109803275343, 0.0, 0.0009644717061794479, 0.0, 0.0, 0.44780560238339745, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.4962679673706645, 0.0, 0.008267299447856608, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.6719286019431624, 0.1932540547332544, 0.6762198255750292, 0.0, 0.0, 0.0003312368464636427, 0.0]	[nan, 0.7085417733756095, 0.8643251797889624, 0.0, 0.0020922282164545967, 0.0006691672739475508, nan, 0.0009725011389865425, 0.0, 0.0, 0.9224475476880146, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.7984415122785299, 0.0, 0.008394275137866055, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.9294223049507054, 0.2306496542338313, 0.7045666997791757, 0.0, 0.0, 0.0003315891206418271, 0.0]

Framework versions

• Transformers 4.19.2
• Pytorch 1.11.0+cu102
• Datasets 2.2.2
• Tokenizers 0.12.1