README.md

metadata

license: other
base_model: nvidia/mit-b0
tags:
  - image-segmentation
  - vision
  - generated_from_trainer
model-index:
  - name: segformer-finetuned-biofilm_MRCNNv1_halfjoin
    results: []

segformer-finetuned-biofilm_MRCNNv1_halfjoin

This model is a fine-tuned version of nvidia/mit-b0 on the heroza/biofilm_MRCNNv1_halfjoin dataset. It achieves the following results on the evaluation set:

• Loss: 0.0208
• Mean Iou: 0.4961
• Mean Accuracy: 0.9923
• Overall Accuracy: 0.9923
• Accuracy Background: 0.9923
• Accuracy Biofilm: nan
• Iou Background: 0.9923
• Iou Biofilm: 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: 8
• eval_batch_size: 8
• seed: 1337
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: polynomial
• training_steps: 10000

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Background	Accuracy Biofilm	Iou Background	Iou Biofilm
0.0713	1.0	478	0.0381	0.4953	0.9906	0.9906	0.9906	nan	0.9906	0.0
0.044	2.0	956	0.0202	0.4975	0.9949	0.9949	0.9949	nan	0.9949	0.0
0.041	3.0	1434	0.0181	0.4972	0.9945	0.9945	0.9945	nan	0.9945	0.0
0.0361	4.0	1912	0.0203	0.4963	0.9926	0.9926	0.9926	nan	0.9926	0.0
0.0357	5.0	2390	0.0163	0.4971	0.9942	0.9942	0.9942	nan	0.9942	0.0
0.0336	6.0	2868	0.0340	0.4958	0.9915	0.9915	0.9915	nan	0.9915	0.0
0.0295	7.0	3346	0.0126	0.4978	0.9955	0.9955	0.9955	nan	0.9955	0.0
0.0251	8.0	3824	0.0220	0.4957	0.9915	0.9915	0.9915	nan	0.9915	0.0
0.0265	9.0	4302	0.0182	0.4966	0.9933	0.9933	0.9933	nan	0.9933	0.0
0.0238	10.0	4780	0.0155	0.4970	0.9940	0.9940	0.9940	nan	0.9940	0.0
0.0258	11.0	5258	0.0181	0.4966	0.9931	0.9931	0.9931	nan	0.9931	0.0
0.0264	12.0	5736	0.0179	0.4969	0.9938	0.9938	0.9938	nan	0.9938	0.0
0.0265	13.0	6214	0.0222	0.4959	0.9917	0.9917	0.9917	nan	0.9917	0.0
0.0219	14.0	6692	0.0200	0.4962	0.9925	0.9925	0.9925	nan	0.9925	0.0
0.0213	15.0	7170	0.0234	0.4958	0.9916	0.9916	0.9916	nan	0.9916	0.0
0.0192	16.0	7648	0.0199	0.4961	0.9922	0.9922	0.9922	nan	0.9922	0.0
0.0232	17.0	8126	0.0208	0.4961	0.9923	0.9923	0.9923	nan	0.9923	0.0
0.0219	18.0	8604	0.0245	0.4955	0.9909	0.9909	0.9909	nan	0.9909	0.0
0.0201	19.0	9082	0.0211	0.4961	0.9922	0.9922	0.9922	nan	0.9922	0.0
0.0192	20.0	9560	0.0207	0.4962	0.9923	0.9923	0.9923	nan	0.9923	0.0
0.0175	20.92	10000	0.0208	0.4961	0.9923	0.9923	0.9923	nan	0.9923	0.0

Framework versions

• Transformers 4.38.0.dev0
• Pytorch 2.0.0+cu117
• Datasets 2.14.4
• Tokenizers 0.15.1