---
license: other
base_model: nvidia/mit-b3
tags:
- generated_from_trainer
model-index:
- name: segformer-roof
  results: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# segformer-roof

This model is a fine-tuned version of [nvidia/mit-b3](https://huggingface.co/nvidia/mit-b3) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1424
- Mean Iou: 0.6570
- Mean Accuracy: 0.7338
- Overall Accuracy: 0.9484
- Per Category Iou: [0.9476783882507867, 0.42934908622698137, 0.5939862114533478]
- Per Category Accuracy: [0.9826387957018243, 0.5398301458913112, 0.6788943731278633]

## 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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10

### Training results

| Training Loss | Epoch | Step | Validation Loss | Mean Iou | Mean Accuracy | Overall Accuracy | Per Category Iou                                              | Per Category Accuracy                                         |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:-------------:|:----------------:|:-------------------------------------------------------------:|:-------------------------------------------------------------:|
| 0.2499        | 1.0   | 930  | 0.1815          | 0.5867   | 0.6662        | 0.9365           | [0.9366364285161356, 0.31719261329477677, 0.5063121008274123] | [0.9792975820886569, 0.391370646753873, 0.6280270152281657]   |
| 0.1757        | 2.0   | 1860 | 0.1714          | 0.6022   | 0.6773        | 0.9395           | [0.9393634573331426, 0.3425379991526554, 0.5248483866382888]  | [0.9812229375223134, 0.428240950515148, 0.6224989524490145]   |
| 0.1594        | 3.0   | 2790 | 0.1629          | 0.6084   | 0.6710        | 0.9420           | [0.9416754823459575, 0.34843136321450174, 0.5350688263682425] | [0.9853655573464773, 0.43172953399691005, 0.5959325404903174] |
| 0.1523        | 4.0   | 3720 | 0.1596          | 0.6076   | 0.6748        | 0.9431           | [0.9431751130845217, 0.3239168742399479, 0.5556505610312258]  | [0.9852091404732521, 0.3708811641866297, 0.6682165559673183]  |
| 0.1439        | 5.0   | 4650 | 0.1523          | 0.6302   | 0.7055        | 0.9446           | [0.9440239408727975, 0.37488313572379306, 0.5716850871490047] | [0.9823225423399531, 0.45512728622828424, 0.6790873748340986] |
| 0.1371        | 6.0   | 5580 | 0.1507          | 0.6435   | 0.7255        | 0.9454           | [0.9448592138832133, 0.40938956878632793, 0.5763842195624918] | [0.9805292644632151, 0.5269108861841263, 0.6691477422245232]  |
| 0.1353        | 7.0   | 6510 | 0.1483          | 0.6535   | 0.7471        | 0.9454           | [0.945006495944485, 0.428182455906198, 0.5872114398937206]    | [0.977372124880732, 0.5920333447023434, 0.6719542751488515]   |
| 0.1313        | 8.0   | 7440 | 0.1475          | 0.6543   | 0.7416        | 0.9465           | [0.9458449483191159, 0.4282098022538787, 0.5888713805135868]  | [0.9792916130278106, 0.5663642070609791, 0.6792639127879244]  |
| 0.1274        | 9.0   | 8370 | 0.1446          | 0.6511   | 0.7257        | 0.9477           | [0.9470057563004056, 0.41812909680083715, 0.5881565577705948] | [0.983095828320569, 0.5191575508625893, 0.6747574079503997]   |
| 0.1274        | 10.0  | 9300 | 0.1424          | 0.6570   | 0.7338        | 0.9484           | [0.9476783882507867, 0.42934908622698137, 0.5939862114533478] | [0.9826387957018243, 0.5398301458913112, 0.6788943731278633]  |


### Framework versions

- Transformers 4.33.1
- Pytorch 2.0.1
- Datasets 2.14.5
- Tokenizers 0.13.3