README.md

metadata

license: other
base_model: nvidia/mit-b0
tags:
  - vision
  - image-segmentation
  - generated_from_trainer
model-index:
  - name: segformer-b0-finetuned-segments-stamp-verification2
    results: []

segformer-b0-finetuned-segments-stamp-verification2

This model is a fine-tuned version of nvidia/mit-b0 on the AliShah07/stamp-verification dataset. It achieves the following results on the evaluation set:

• Loss: 0.0365
• Mean Iou: 0.1372
• Mean Accuracy: 0.2744
• Overall Accuracy: 0.2744
• Accuracy Unlabeled: nan
• Accuracy Stamp: 0.2744
• Iou Unlabeled: 0.0
• Iou Stamp: 0.2744

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: 2
• eval_batch_size: 2
• 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 Unlabeled	Accuracy Stamp	Iou Unlabeled	Iou Stamp
0.4566	0.8333	20	0.4738	0.1430	0.2860	0.2860	nan	0.2860	0.0	0.2860
0.3076	1.6667	40	0.3046	0.1307	0.2614	0.2614	nan	0.2614	0.0	0.2614
0.2373	2.5	60	0.2226	0.0604	0.1209	0.1209	nan	0.1209	0.0	0.1209
0.2184	3.3333	80	0.2220	0.1942	0.3884	0.3884	nan	0.3884	0.0	0.3884
0.1578	4.1667	100	0.1704	0.2468	0.4936	0.4936	nan	0.4936	0.0	0.4936
0.1412	5.0	120	0.1269	0.0376	0.0751	0.0751	nan	0.0751	0.0	0.0751
0.1109	5.8333	140	0.1076	0.2741	0.5483	0.5483	nan	0.5483	0.0	0.5483
0.106	6.6667	160	0.0892	0.0583	0.1166	0.1166	nan	0.1166	0.0	0.1166
0.0899	7.5	180	0.0747	0.0173	0.0346	0.0346	nan	0.0346	0.0	0.0346
0.0794	8.3333	200	0.0683	0.0189	0.0378	0.0378	nan	0.0378	0.0	0.0378
0.0741	9.1667	220	0.0639	0.0981	0.1963	0.1963	nan	0.1963	0.0	0.1963
0.0832	10.0	240	0.0559	0.0599	0.1198	0.1198	nan	0.1198	0.0	0.1198
0.0575	10.8333	260	0.0527	0.0769	0.1538	0.1538	nan	0.1538	0.0	0.1538
0.05	11.6667	280	0.0502	0.0852	0.1704	0.1704	nan	0.1704	0.0	0.1704
0.0523	12.5	300	0.0446	0.1038	0.2076	0.2076	nan	0.2076	0.0	0.2076
0.0481	13.3333	320	0.0431	0.0956	0.1913	0.1913	nan	0.1913	0.0	0.1913
0.0471	14.1667	340	0.0420	0.1330	0.2660	0.2660	nan	0.2660	0.0	0.2660
0.042	15.0	360	0.0412	0.1124	0.2248	0.2248	nan	0.2248	0.0	0.2248
0.041	15.8333	380	0.0400	0.1144	0.2288	0.2288	nan	0.2288	0.0	0.2288
0.0444	16.6667	400	0.0383	0.1415	0.2830	0.2830	nan	0.2830	0.0	0.2830
0.0514	17.5	420	0.0377	0.0779	0.1559	0.1559	nan	0.1559	0.0	0.1559
0.0434	18.3333	440	0.0374	0.1482	0.2964	0.2964	nan	0.2964	0.0	0.2964
0.0383	19.1667	460	0.0363	0.1843	0.3686	0.3686	nan	0.3686	0.0	0.3686
0.0411	20.0	480	0.0365	0.1372	0.2744	0.2744	nan	0.2744	0.0	0.2744

Framework versions

• Transformers 4.40.2
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1