segformer-b0_DsApollo_ILT

This model is a fine-tuned version of ironchanchellor/segformer-b0_DsA on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0116
• Mean Iou: 0.7963
• Mean Accuracy: 0.9731
• Overall Accuracy: 0.9942
• Accuracy Background: nan
• Accuracy Haz: 0.9965
• Accuracy Matrix: 0.9883
• Accuracy Porosity: 0.9077
• Accuracy Carbides: 0.9757
• Accuracy Substrate: 0.9970
• Iou Background: 0.0
• Iou Haz: 0.9932
• Iou Matrix: 0.9767
• Iou Porosity: 0.8590
• Iou Carbides: 0.9547
• Iou Substrate: 0.9942

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Background	Accuracy Haz	Accuracy Matrix	Accuracy Porosity	Accuracy Carbides	Accuracy Substrate	Iou Background	Iou Haz	Iou Matrix	Iou Porosity	Iou Carbides	Iou Substrate
0.2652	1.0	591	0.0181	0.7893	0.9731	0.9914	nan	0.9952	0.9836	0.9148	0.9794	0.9926	0.0	0.9872	0.9741	0.8340	0.9516	0.9886
0.0167	2.0	1182	0.0196	0.7899	0.9662	0.9920	nan	0.9919	0.9849	0.8780	0.9793	0.9967	0.0	0.9884	0.9746	0.8346	0.9522	0.9899
0.0272	3.0	1773	0.0152	0.7895	0.9652	0.9923	nan	0.9970	0.9906	0.8787	0.9672	0.9926	0.0	0.9890	0.9753	0.8311	0.9512	0.9903
0.0117	4.0	2364	0.0131	0.7924	0.9663	0.9933	nan	0.9951	0.9902	0.8801	0.9698	0.9966	0.0	0.9912	0.9759	0.8425	0.9524	0.9924
0.0176	5.0	2955	0.0160	0.7919	0.9658	0.9930	nan	0.9939	0.9871	0.8731	0.9780	0.9968	0.0	0.9903	0.9760	0.8396	0.9536	0.9916
0.0772	6.0	3546	0.0125	0.7928	0.9671	0.9937	nan	0.9949	0.9858	0.8767	0.9804	0.9976	0.0	0.9922	0.9760	0.8419	0.9535	0.9933
0.0568	7.0	4137	0.0126	0.7950	0.9697	0.9937	nan	0.9978	0.9895	0.8933	0.9735	0.9946	0.0	0.9920	0.9766	0.8540	0.9540	0.9931
0.0228	8.0	4728	0.0158	0.7956	0.9766	0.9934	nan	0.9933	0.9882	0.9283	0.9749	0.9980	0.0	0.9911	0.9766	0.8591	0.9543	0.9923
0.0058	9.0	5319	0.0120	0.7960	0.9731	0.9940	nan	0.9958	0.9873	0.9068	0.9782	0.9972	0.0	0.9927	0.9766	0.8580	0.9547	0.9938
0.0038	10.0	5910	0.0116	0.7963	0.9731	0.9942	nan	0.9965	0.9883	0.9077	0.9757	0.9970	0.0	0.9932	0.9767	0.8590	0.9547	0.9942

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0