---
license: other
base_model: nvidia/mit-b5
tags:
- vision
- image-segmentation
- generated_from_trainer
model-index:
- name: SegFormer_Clean_Set1_95images_mit-b5
  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_Clean_Set1_95images_mit-b5_Grayscale

This model is a fine-tuned version of [nvidia/mit-b5](https://huggingface.co/nvidia/mit-b5) on the Hasano20/Clean_Set1_95images dataset.
It achieves the following results on the evaluation set:
- Train Loss:  0.0103
- Val Loss: 0.0229
- Mean Iou: 0.9729
- Mean Accuracy: 0.9859
- Overall Accuracy: 0.9928
- Accuracy Background: 0.9972
- Accuracy Melt: 0.9669
- Accuracy Substrate: 0.9937
- Iou Background: 0.9944
- Iou Melt: 0.9370
- Iou Substrate: 0.9871

## 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: 0.0001
- 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: cosine
- lr_scheduler_warmup_steps: 100
- num_epochs: 50

### Training results

| Training Loss | Epoch   | Step | Validation Loss | Mean Iou | Mean Accuracy | Overall Accuracy | Accuracy Background | Accuracy Melt | Accuracy Substrate | Iou Background | Iou Melt | Iou Substrate |
|:-------------:|:-------:|:----:|:---------------:|:--------:|:-------------:|:----------------:|:-------------------:|:-------------:|:------------------:|:--------------:|:--------:|:-------------:|
| 0.1375        | 5.5556  | 50   | 0.1577          | 0.7820   | 0.8338        | 0.9411           | 0.9857              | 0.5352        | 0.9806             | 0.9746         | 0.4754   | 0.8959        |
| 0.0403        | 11.1111 | 100  | 0.1948          | 0.7535   | 0.7960        | 0.9378           | 0.9893              | 0.4011        | 0.9977             | 0.9826         | 0.3954   | 0.8825        |
| 0.0291        | 16.6667 | 150  | 0.0484          | 0.9337   | 0.9479        | 0.9832           | 0.9969              | 0.8495        | 0.9973             | 0.9884         | 0.8414   | 0.9712        |
| 0.0114        | 22.2222 | 200  | 0.0273          | 0.9634   | 0.9808        | 0.9903           | 0.9930              | 0.9544        | 0.9950             | 0.9917         | 0.9149   | 0.9838        |
| 0.0138        | 27.7778 | 250  | 0.0289          | 0.9655   | 0.9782        | 0.9910           | 0.9966              | 0.9423        | 0.9956             | 0.9941         | 0.9190   | 0.9836        |
| 0.0072        | 33.3333 | 300  | 0.0257          | 0.9689   | 0.9855        | 0.9918           | 0.9975              | 0.9682        | 0.9908             | 0.9945         | 0.9276   | 0.9847        |
| 0.007         | 38.8889 | 350  | 0.0234          | 0.9722   | 0.9862        | 0.9926           | 0.9968              | 0.9684        | 0.9934             | 0.9944         | 0.9354   | 0.9867        |
| 0.0063        | 44.4444 | 400  | 0.0232          | 0.9727   | 0.9866        | 0.9927           | 0.9971              | 0.9696        | 0.9931             | 0.9945         | 0.9366   | 0.9870        |
| 0.0103        | 50.0    | 450  | 0.0229          | 0.9729   | 0.9859        | 0.9928           | 0.9972              | 0.9669        | 0.9937             | 0.9944         | 0.9370   | 0.9871        |


### Framework versions

- Transformers 4.41.2
- Pytorch 2.0.1+cu117
- Datasets 2.19.2
- Tokenizers 0.19.1