ModelPrésentéExterne

Autocrop Bilder

This model is a fine-tuned version of nvidia/mit-b0 on the /mnt/disk1/autocrop-data/datasets/bilder dataset. It achieves the following results on the evaluation set: Loss: 0.0434 Mean Iou: 0.4950 Mean Accuracy:...

Sourcehuggingface.co/NbAiLab/autocrop-bilder/tree/3e343901f756ce9d8e90a26e3507c2fd047d7757

TypeOtherModèle de basenvidia/mit-b0VersionRevision 3e34390CClasse

Source du modèle

Type: Other
Taille du fichier: 14,2 MB
Modèle de base: nvidia/mit-b0
Version: vRevision 3e34390
Mis à jour: 16 sept. 2026
Téléchargements: 6
Source: Hugging Face

Description de la source

autocrop-bilder

This model is a fine-tuned version of nvidia/mit-b0 on the /mnt/disk1/autocrop-data/datasets/bilder dataset. It achieves the following results on the evaluation set:

Loss: 0.0434
Mean Iou: 0.4950
Mean Accuracy: 0.9899
Overall Accuracy: 0.9899
Accuracy Background: nan
Accuracy Crop: 0.9899
Iou Background: 0.0
Iou Crop: 0.9899

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: 42
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 0.1
num_epochs: 50.0
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Background	Accuracy Crop	Iou Crop
0.3277	1.0	112	0.3304	0.4474	0.8948	0.8948	nan	0.8948	0.8948
0.1834	2.0	224	0.1733	0.4725	0.9450	0.9450	nan	0.9450	0.9450
0.1279	3.0	336	0.1177	0.4907	0.9813	0.9813	nan	0.9813	0.9813
0.0879	4.0	448	0.0841	0.4929	0.9858	0.9858	nan	0.9858	0.9858
0.0796	5.0	560	0.0840	0.4871	0.9742	0.9742	nan	0.9742	0.9742
0.0641	6.0	672	0.0709	0.4930	0.9860	0.9860	nan	0.9860	0.9860
0.0523	7.0	784	0.0633	0.4947	0.9894	0.9894	nan	0.9894	0.9894
0.0577	8.0	896	0.0606	0.4904	0.9807	0.9807	nan	0.9807	0.9807
0.0528	9.0	1008	0.0596	0.4952	0.9904	0.9904	nan	0.9904	0.9904
0.0449	10.0	1120	0.0565	0.4925	0.9850	0.9850	nan	0.9850	0.9850
0.0466	11.0	1232	0.0533	0.4926	0.9853	0.9853	nan	0.9853	0.9853
0.0464	12.0	1344	0.0500	0.4937	0.9874	0.9874	nan	0.9874	0.9874
0.0456	13.0	1456	0.0503	0.4957	0.9914	0.9914	nan	0.9914	0.9914
0.0394	14.0	1568	0.0491	0.4938	0.9876	0.9876	nan	0.9876	0.9876
0.0402	15.0	1680	0.0514	0.4960	0.9921	0.9921	nan	0.9921	0.9921
0.0421	16.0	1792	0.0489	0.4955	0.9910	0.9910	nan	0.9910	0.9910
0.0453	17.0	1904	0.0461	0.4947	0.9894	0.9894	nan	0.9894	0.9894
0.0449	18.0	2016	0.0485	0.4929	0.9858	0.9858	nan	0.9858	0.9858
0.0349	19.0	2128	0.0468	0.4962	0.9925	0.9925	nan	0.9925	0.9925
0.0351	20.0	2240	0.0470	0.4962	0.9924	0.9924	nan	0.9924	0.9924
0.0324	21.0	2352	0.0452	0.4949	0.9897	0.9897	nan	0.9897	0.9897
0.0367	22.0	2464	0.0461	0.4949	0.9897	0.9897	nan	0.9897	0.9897
0.0350	23.0	2576	0.0451	0.4952	0.9903	0.9903	nan	0.9903	0.9903
0.0354	24.0	2688	0.0469	0.4957	0.9914	0.9914	nan	0.9914	0.9914
0.0353	25.0	2800	0.0452	0.4945	0.9890	0.9890	nan	0.9890	0.9890
0.0334	26.0	2912	0.0448	0.4962	0.9924	0.9924	nan	0.9924	0.9924
0.0269	27.0	3024	0.0448	0.4958	0.9915	0.9915	nan	0.9915	0.9915
0.0319	28.0	3136	0.0443	0.4949	0.9898	0.9898	nan	0.9898	0.9898
0.0293	29.0	3248	0.0450	0.4962	0.9924	0.9924	nan	0.9924	0.9924
0.0306	30.0	3360	0.0438	0.4962	0.9923	0.9923	nan	0.9923	0.9923
0.0278	31.0	3472	0.0447	0.4960	0.9920	0.9920	nan	0.9920	0.9920
0.0268	32.0	3584	0.0459	0.4962	0.9924	0.9924	nan	0.9924	0.9924
0.0269	33.0	3696	0.0434	0.4950	0.9899	0.9899	nan	0.9899	0.9899
0.0268	34.0	3808	0.0445	0.4953	0.9906	0.9906	nan	0.9906	0.9906
0.0302	35.0	3920	0.0443	0.4946	0.9891	0.9891	nan	0.9891	0.9891
0.0239	36.0	4032	0.0439	0.4959	0.9919	0.9919	nan	0.9919	0.9919
0.0268	37.0	4144	0.0442	0.4958	0.9915	0.9915	nan	0.9915	0.9915
0.0318	38.0	4256	0.0451	0.4958	0.9916	0.9916	nan	0.9916	0.9916
0.0276	39.0	4368	0.0444	0.4956	0.9912	0.9912	nan	0.9912	0.9912
0.0248	40.0	4480	0.0456	0.4960	0.9921	0.9921	nan	0.9921	0.9921
0.0244	41.0	4592	0.0449	0.4952	0.9905	0.9905	nan	0.9905	0.9905
0.0235	42.0	4704	0.0445	0.4961	0.9922	0.9922	nan	0.9922	0.9922
0.0241	43.0	4816	0.0445	0.4960	0.9920	0.9920	nan	0.9920	0.9920
0.0295	44.0	4928	0.0445	0.4959	0.9919	0.9919	nan	0.9919	0.9919
0.0252	45.0	5040	0.0443	0.4960	0.9919	0.9919	nan	0.9919	0.9919
0.0213	46.0	5152	0.0443	0.4961	0.9922	0.9922	nan	0.9922	0.9922
0.0238	47.0	5264	0.0446	0.4958	0.9917	0.9917	nan	0.9917	0.9917
0.0234	48.0	5376	0.0445	0.4959	0.9918	0.9918	nan	0.9918	0.9918
0.0223	49.0	5488	0.0445	0.4957	0.9914	0.9914	nan	0.9914	0.9914
0.0245	50.0	5600	0.0447	0.4959	0.9917	0.9917	nan	0.9917	0.9917

Framework versions

Transformers 5.8.0
Pytorch 2.11.0+cu130
Datasets 4.8.5
Tokenizers 0.22.2

Sources

1 source

Hugging FacePrincipalVersion 3e343901f756ce9d8e90a26e3507c2fd047d7757

huggingface.co/NbAiLab/autocrop-bilder/tree/3e343901f756ce9d8e90a26e3507c2fd047d7757

Vérifié 1 sept.

Artefacts du modèle

1 artefact

model.safetensors

safetensors · 14,2 MB · SHA-256 e1023e5513fc…0ab1 · Hugging Face

Download file

Extraits de sources

2 extraits

Readme1 sept.huggingface.co/NbAiLab/autocrop-bilder

--- library_name: transformers license: other base_model: nvidia/mit-b0 tags: - image-segmentation - vision - generated_from_trainer datasets: - generator model-index: - name: autocrop-bilder results: [] ---  # autocrop-bilder This model is a fine-tuned version of [nvidia/mit-b0](https://huggingface.co/nvidia/mit-b0) on the /mnt/disk1/autocrop-data/datasets/bilder dataset. It achieves the following results on the evaluation set: - Loss: 0.0434 - Mean Iou: 0.4950 - Mean Accuracy: 0.9899 - Overall Accuracy: 0.9899 - Accuracy Background: nan - Accuracy Crop: 0.9899 - Iou Background: 0.0 - Iou Crop: 0.9899 ## 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: 42 - optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 0.1 - num_epochs: 50.0 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Mean Iou | Mean Accuracy | Overall Accuracy | Accuracy Background | Accuracy Crop | Iou Background | Iou Crop | |:-------------:|:-----:|:----:|:---------------:|:--------:|:-------------:|:----------------:|:-------------------:|:-------------:|:--------------:|:--------:| | 0.3277 | 1.0 | 112 | 0.3304 | 0.4474 | 0.8948 | 0.8948 | nan | 0.8948 | 0.0 | 0.8948 | | 0.1834 | 2.0 | 224 | 0.1733 | 0.4725 | 0.9450 | 0.9450 | nan | 0.9450 | 0.0 | 0.9450 | | 0.1279 | 3.0 | 336 | 0.1177 | 0.4907 | 0.9813 | 0.9813 | nan | 0.9813 | 0.0 | 0.9813 | | 0.0879 | 4.0 | 448 | 0.0841 | 0.4929 | 0.9858 | 0.9858 | nan | 0.9858 | 0.0 | 0.9858 | | 0.0796 | 5.0 | 560 | 0.0840 | 0.4871 | 0.9742 | 0.9742 | nan | 0.9742 | 0.0 | 0.9742 | | 0.0641 | 6.0 | 672 | 0.0709 | 0.4930 | 0.9860 | 0.9860 | nan | 0.9860 | 0.0 | 0.9860 | | 0.0523 | 7.0 | 784 | 0.0633 | 0.4947 | 0.9894 | 0.9894 | nan | 0.9894 | 0.0 | 0.9894 | | 0.0577 | 8...

Note1 sept.huggingface.co/NbAiLab/autocrop-bilder

Source context: 6 downloads · 0 likes · Pipeline image-segmentation · Library transformers · Repo NbAiLab/autocrop-bilder