ModelPrésentéExterne

Autocrop Combined

This model is a fine-tuned version of nvidia/mit-b0 on the /mnt/disk1/autocrop-data/datasets/combined dataset. It achieves the following results on the evaluation set: Loss: 0.0306 Mean Iou: 0.4956 Mean Accuracy:...

Sourcehuggingface.co/NbAiLab/autocrop-combined/tree/af5c9792cff61d5b77415aadc1a91bd15afdd161

TypeOtherModèle de basenvidia/mit-b0VersionRevision af5c979CClasse

Source du modèle

Type: Other
Taille du fichier: 14,2 MB
Modèle de base: nvidia/mit-b0
Version: vRevision af5c979
Mis à jour: 16 sept. 2026
Téléchargements: 6
Source: Hugging Face

Description de la source

autocrop-combined

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

Loss: 0.0306
Mean Iou: 0.4956
Mean Accuracy: 0.9911
Overall Accuracy: 0.9911
Accuracy Background: nan
Accuracy Crop: 0.9911
Iou Background: 0.0
Iou Crop: 0.9911

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
0.1895	1.0	328	0.1511	0.4885	0.9771	0.9771	nan	0.9771	0.9771
0.0960	2.0	656	0.0832	0.4822	0.9644	0.9644	nan	0.9644	0.9644
0.0912	3.0	984	0.0641	0.4927	0.9854	0.9854	nan	0.9854	0.9854
0.0654	4.0	1312	0.0554	0.4935	0.9870	0.9870	nan	0.9870	0.9870
0.0499	5.0	1640	0.0455	0.4920	0.9841	0.9841	nan	0.9841	0.9841
0.0598	6.0	1968	0.0433	0.4930	0.9861	0.9861	nan	0.9861	0.9861
0.0705	7.0	2296	0.0418	0.4910	0.9820	0.9820	nan	0.9820	0.9820
0.0425	8.0	2624	0.0398	0.4935	0.9869	0.9869	nan	0.9869	0.9869
0.0352	9.0	2952	0.0392	0.4939	0.9878	0.9878	nan	0.9878	0.9878
0.0430	10.0	3280	0.0389	0.4917	0.9834	0.9834	nan	0.9834	0.9834
0.0478	11.0	3608	0.0392	0.4916	0.9832	0.9832	nan	0.9832	0.9832
0.0540	12.0	3936	0.0390	0.4947	0.9894	0.9894	nan	0.9894	0.9894
0.0401	13.0	4264	0.0357	0.4935	0.9870	0.9870	nan	0.9870	0.9870
0.0391	14.0	4592	0.0352	0.4951	0.9903	0.9903	nan	0.9903	0.9903
0.0295	15.0	4920	0.0353	0.4941	0.9882	0.9882	nan	0.9882	0.9882
0.0328	16.0	5248	0.0352	0.4931	0.9863	0.9863	nan	0.9863	0.9863
0.0297	17.0	5576	0.0353	0.4939	0.9878	0.9878	nan	0.9878	0.9878
0.0253	18.0	5904	0.0351	0.4939	0.9878	0.9878	nan	0.9878	0.9878
0.0317	19.0	6232	0.0399	0.4957	0.9915	0.9915	nan	0.9915	0.9915
0.0258	20.0	6560	0.0339	0.4936	0.9873	0.9873	nan	0.9873	0.9873
0.0300	21.0	6888	0.0343	0.4934	0.9868	0.9868	nan	0.9868	0.9868
0.0312	22.0	7216	0.0364	0.4957	0.9915	0.9915	nan	0.9915	0.9915
0.0256	23.0	7544	0.0329	0.4942	0.9884	0.9884	nan	0.9884	0.9884
0.0238	24.0	7872	0.0322	0.4946	0.9893	0.9893	nan	0.9893	0.9893
0.0238	25.0	8200	0.0325	0.4940	0.9880	0.9880	nan	0.9880	0.9880
0.0233	26.0	8528	0.0328	0.4949	0.9898	0.9898	nan	0.9898	0.9898
0.0239	27.0	8856	0.0323	0.4951	0.9903	0.9903	nan	0.9903	0.9903
0.0258	28.0	9184	0.0332	0.4949	0.9898	0.9898	nan	0.9898	0.9898
0.0235	29.0	9512	0.0325	0.4949	0.9897	0.9897	nan	0.9897	0.9897
0.0226	30.0	9840	0.0331	0.4955	0.9911	0.9911	nan	0.9911	0.9911
0.0216	31.0	10168	0.0322	0.4957	0.9915	0.9915	nan	0.9915	0.9915
0.0304	32.0	10496	0.0318	0.4952	0.9903	0.9903	nan	0.9903	0.9903
0.0210	33.0	10824	0.0313	0.4952	0.9904	0.9904	nan	0.9904	0.9904
0.0200	34.0	11152	0.0321	0.4962	0.9923	0.9923	nan	0.9923	0.9923
0.0237	35.0	11480	0.0318	0.4956	0.9913	0.9913	nan	0.9913	0.9913
0.0320	36.0	11808	0.0316	0.4938	0.9876	0.9876	nan	0.9876	0.9876
0.0225	37.0	12136	0.0315	0.4941	0.9882	0.9882	nan	0.9882	0.9882
0.0208	38.0	12464	0.0306	0.4956	0.9911	0.9911	nan	0.9911	0.9911
0.0219	39.0	12792	0.0313	0.4950	0.9900	0.9900	nan	0.9900	0.9900
0.0222	40.0	13120	0.0311	0.4953	0.9906	0.9906	nan	0.9906	0.9906
0.0190	41.0	13448	0.0310	0.4954	0.9907	0.9907	nan	0.9907	0.9907
0.0200	42.0	13776	0.0308	0.4951	0.9902	0.9902	nan	0.9902	0.9902
0.0209	43.0	14104	0.0312	0.4951	0.9903	0.9903	nan	0.9903	0.9903
0.0224	44.0	14432	0.0312	0.4953	0.9905	0.9905	nan	0.9905	0.9905
0.0266	45.0	14760	0.0307	0.4955	0.9909	0.9909	nan	0.9909	0.9909
0.0173	46.0	15088	0.0310	0.4952	0.9904	0.9904	nan	0.9904	0.9904
0.0191	47.0	15416	0.0312	0.4955	0.9910	0.9910	nan	0.9910	0.9910
0.0218	48.0	15744	0.0311	0.4953	0.9905	0.9905	nan	0.9905	0.9905
0.0204	49.0	16072	0.0311	0.4953	0.9905	0.9905	nan	0.9905	0.9905
0.0174	50.0	16400	0.0311	0.4954	0.9907	0.9907	nan	0.9907	0.9907

Framework versions

Transformers 5.8.0
Pytorch 2.11.0+cu130
Datasets 4.8.5
Tokenizers 0.22.2

Sources

1 source

Hugging FacePrincipalVersion af5c9792cff61d5b77415aadc1a91bd15afdd161

huggingface.co/NbAiLab/autocrop-combined/tree/af5c9792cff61d5b77415aadc1a91bd15afdd161

Vérifié 1 sept.

Artefacts du modèle

1 artefact

model.safetensors

safetensors · 14,2 MB · SHA-256 d9c388f5fd42…3c5f · Hugging Face

Download file

Extraits de sources

2 extraits

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

--- library_name: transformers license: other base_model: nvidia/mit-b0 tags: - image-segmentation - vision - generated_from_trainer model-index: - name: autocrop-combined results: [] ---  # autocrop-combined This model is a fine-tuned version of [nvidia/mit-b0](https://huggingface.co/nvidia/mit-b0) on the /mnt/disk1/autocrop-data/datasets/combined dataset. It achieves the following results on the evaluation set: - Loss: 0.0306 - Mean Iou: 0.4956 - Mean Accuracy: 0.9911 - Overall Accuracy: 0.9911 - Accuracy Background: nan - Accuracy Crop: 0.9911 - Iou Background: 0.0 - Iou Crop: 0.9911 ## 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 | |:-------------:|:-----:|:-----:|:---------------:|:--------:|:-------------:|:----------------:|:-------------------:|:-------------:|:--------------:|:--------:| | 0.1895 | 1.0 | 328 | 0.1511 | 0.4885 | 0.9771 | 0.9771 | nan | 0.9771 | 0.0 | 0.9771 | | 0.0960 | 2.0 | 656 | 0.0832 | 0.4822 | 0.9644 | 0.9644 | nan | 0.9644 | 0.0 | 0.9644 | | 0.0912 | 3.0 | 984 | 0.0641 | 0.4927 | 0.9854 | 0.9854 | nan | 0.9854 | 0.0 | 0.9854 | | 0.0654 | 4.0 | 1312 | 0.0554 | 0.4935 | 0.9870 | 0.9870 | nan | 0.9870 | 0.0 | 0.9870 | | 0.0499 | 5.0 | 1640 | 0.0455 | 0.4920 | 0.9841 | 0.9841 | nan | 0.9841 | 0.0 | 0.9841 | | 0.0598 | 6.0 | 1968 | 0.0433 | 0.4930 | 0.9861 | 0.9861 | nan | 0.9861 | 0.0 | 0.9861 | | 0.0705 | 7.0 | 2296 | 0.0418 | 0.4910 | 0.9820 | 0.9820 | nan | 0.9820 | 0.0 | 0.9820 | | 0.0425 | 8.0 | 2624 |...

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

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