test-segformer-1

This model is a fine-tuned version of nvidia/mit-b0 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5184
• Mean Iou: 0.5013
• Mean Accuracy: 0.6080
• Overall Accuracy: 0.8860
• Accuracy Background: 0.9710
• Accuracy Aeroplane: 0.7936
• Accuracy Bicycle: 0.6422
• Accuracy Bird: 0.6118
• Accuracy Boat: 0.4959
• Accuracy Bottle: 0.3990
• Accuracy Bus: 0.7997
• Accuracy Car: 0.7558
• Accuracy Cat: 0.8353
• Accuracy Chair: 0.3141
• Accuracy Cow: 0.4212
• Accuracy Diningtable: 0.2675
• Accuracy Dog: 0.5355
• Accuracy Horse: 0.5817
• Accuracy Motorbike: 0.7442
• Accuracy Person: 0.8677
• Accuracy Pottedplant: 0.3351
• Accuracy Sheep: 0.7845
• Accuracy Sofa: 0.3684
• Accuracy Train: 0.6319
• Accuracy Tvmonitor: 0.6111
• Iou Background: 0.9054
• Iou Aeroplane: 0.7364
• Iou Bicycle: 0.4610
• Iou Bird: 0.4835
• Iou Boat: 0.4281
• Iou Bottle: 0.3470
• Iou Bus: 0.7430
• Iou Car: 0.6847
• Iou Cat: 0.6004
• Iou Chair: 0.2188
• Iou Cow: 0.3697
• Iou Diningtable: 0.2446
• Iou Dog: 0.4173
• Iou Horse: 0.3859
• Iou Motorbike: 0.5811
• Iou Person: 0.7178
• Iou Pottedplant: 0.3016
• Iou Sheep: 0.5628
• Iou Sofa: 0.2973
• Iou Train: 0.5734
• Iou Tvmonitor: 0.4668

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: 0.001
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 99

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Background	Accuracy Aeroplane	Accuracy Bicycle	Accuracy Bird	Accuracy Boat	Accuracy Bottle	Accuracy Bus	Accuracy Car	Accuracy Cat	Accuracy Chair	Accuracy Cow	Accuracy Diningtable	Accuracy Dog	Accuracy Horse	Accuracy Motorbike	Accuracy Person	Accuracy Pottedplant	Accuracy Sheep	Accuracy Sofa	Accuracy Train	Accuracy Tvmonitor	Iou Background	Iou Aeroplane	Iou Bicycle	Iou Bird	Iou Boat	Iou Bottle	Iou Bus	Iou Car	Iou Cat	Iou Chair	Iou Cow	Iou Diningtable	Iou Dog	Iou Horse	Iou Motorbike	Iou Person	Iou Pottedplant	Iou Sheep	Iou Sofa	Iou Train	Iou Tvmonitor
0.3686	21.7391	1000	0.5845	0.3211	0.4345	0.8352	0.9694	0.6936	0.5146	0.4319	0.4197	0.1345	0.3523	0.2277	0.6193	0.0743	0.3442	0.4724	0.5060	0.2571	0.5776	0.7439	0.2928	0.7685	0.2055	0.3511	0.1672	0.8669	0.5458	0.2152	0.3433	0.2435	0.1265	0.3481	0.1988	0.5060	0.0624	0.2761	0.3396	0.3731	0.2182	0.3127	0.6209	0.1716	0.3419	0.1634	0.3243	0.1455
0.2146	43.4783	2000	0.5172	0.4095	0.5204	0.8595	0.9684	0.6709	0.6059	0.4971	0.4196	0.2827	0.5828	0.3871	0.7104	0.2592	0.5440	0.2338	0.4815	0.3928	0.4747	0.8668	0.4723	0.7557	0.1945	0.5497	0.5785	0.8846	0.6450	0.3218	0.3542	0.3466	0.2658	0.4982	0.3832	0.5423	0.1800	0.4389	0.2254	0.3663	0.3070	0.3989	0.6469	0.2971	0.4598	0.1728	0.4736	0.3907
0.0919	65.2174	3000	0.5242	0.4735	0.5733	0.8780	0.9721	0.7582	0.5962	0.5413	0.4951	0.3764	0.8138	0.7061	0.8332	0.2066	0.2866	0.3386	0.5985	0.3335	0.6679	0.8557	0.4816	0.6698	0.2637	0.5505	0.6939	0.8981	0.7151	0.4401	0.4610	0.3947	0.3450	0.7176	0.6253	0.6174	0.1626	0.2620	0.2859	0.3921	0.2445	0.5372	0.7023	0.3574	0.5465	0.2268	0.5270	0.4854
0.0775	86.9565	4000	0.5184	0.5013	0.6080	0.8860	0.9710	0.7936	0.6422	0.6118	0.4959	0.3990	0.7997	0.7558	0.8353	0.3141	0.4212	0.2675	0.5355	0.5817	0.7442	0.8677	0.3351	0.7845	0.3684	0.6319	0.6111	0.9054	0.7364	0.4610	0.4835	0.4281	0.3470	0.7430	0.6847	0.6004	0.2188	0.3697	0.2446	0.4173	0.3859	0.5811	0.7178	0.3016	0.5628	0.2973	0.5734	0.4668

Framework versions

• Transformers 4.44.2
• Pytorch 2.3.0
• Datasets 3.0.0
• Tokenizers 0.19.1