segformer-b0-finetuned-100by100PNG-50epochs

This model is a fine-tuned version of nvidia/mit-b0 on the JCAI2000/100By100BranchPNG dataset. It achieves the following results on the evaluation set:

• Loss: 0.1488
• Mean Iou: 0.0
• Mean Accuracy: nan
• Overall Accuracy: nan
• Accuracy 0: nan
• Iou 0: 0.0

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: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy 0	Iou 0
0.5785	2.63	50	0.4624	0.0	nan	nan	nan	0.0
0.3842	5.26	100	0.3092	0.0	nan	nan	nan	0.0
0.3916	7.89	150	0.2549	0.0	nan	nan	nan	0.0
0.2686	10.53	200	0.2442	0.0	nan	nan	nan	0.0
0.163	13.16	250	0.2046	0.0	nan	nan	nan	0.0
0.118	15.79	300	0.1911	0.0	nan	nan	nan	0.0
0.1432	18.42	350	0.1718	0.0	nan	nan	nan	0.0
0.1692	21.05	400	0.1877	0.0	nan	nan	nan	0.0
0.1444	23.68	450	0.1637	0.0	nan	nan	nan	0.0
0.084	26.32	500	0.1649	0.0	nan	nan	nan	0.0
0.1204	28.95	550	0.1601	0.0	nan	nan	nan	0.0
0.0816	31.58	600	0.1560	0.0	nan	nan	nan	0.0
0.1041	34.21	650	0.1618	0.0	nan	nan	nan	0.0
0.0906	36.84	700	0.1565	0.0	nan	nan	nan	0.0
0.0644	39.47	750	0.1500	0.0	nan	nan	nan	0.0
0.0774	42.11	800	0.1552	0.0	nan	nan	nan	0.0
0.0687	44.74	850	0.1504	0.0	nan	nan	nan	0.0
0.0931	47.37	900	0.1465	0.0	nan	nan	nan	0.0
0.0852	50.0	950	0.1488	0.0	nan	nan	nan	0.0

Framework versions

• Transformers 4.33.0
• Pytorch 2.0.1+cu117
• Datasets 2.14.4
• Tokenizers 0.13.3