segformer-b0-scene-parse-150-lr-5-e-15

This model is a fine-tuned version of DiTo97/binarization-segformer-b3 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2657
• Mean Iou: 0.4845
• Mean Accuracy: 0.5001
• Overall Accuracy: 0.9672
• Per Category Iou: [0.0018194025597222916, 0.9671517415294609]
• Per Category Accuracy: [0.001918102131300032, 0.9982521972361976]

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: 1e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Per Category Iou	Per Category Accuracy
No log	1.0	112	2.2288	0.0208	0.4868	0.0410	[0.03036636790421317, 0.011265774627153866]	[0.9622473367236779, 0.011279477594064372]
No log	2.0	224	1.6154	0.0182	0.4963	0.0362	[0.03097736523913424, 0.005513312495278201]	[0.9871504131558042, 0.005515594979208372]
No log	3.0	336	0.9216	0.1937	0.5158	0.3688	[0.032185306965168796, 0.3552501717296959]	[0.672525648250623, 0.3589983267382158]
No log	4.0	448	0.9276	0.1561	0.5134	0.2969	[0.03198740212709094, 0.280147471502915]	[0.7443848833182828, 0.28245463938025656]
1.4322	5.0	560	0.6011	0.4362	0.5033	0.8459	[0.0271617976460957, 0.8452071385383193]	[0.13786740991709726, 0.8686841695959868]
1.4322	6.0	672	0.3566	0.4843	0.4999	0.9653	[0.003156516583524233, 0.9653443351384307]	[0.0035153889503737753, 0.9963369917295061]
1.4322	7.0	784	0.4510	0.4833	0.5026	0.9515	[0.015110478622284323, 0.9514896636755739]	[0.023826902315981016, 0.981414850138177]
1.4322	8.0	896	0.3993	0.4862	0.5025	0.9626	[0.009768906238396621, 0.9625427377471698]	[0.011834520406569755, 0.9931874576024252]
0.4808	9.0	1008	0.3568	0.4846	0.5002	0.9663	[0.002888368095508705, 0.9662512532108187]	[0.003131768524113769, 0.9972849692353025]
0.4808	10.0	1120	0.3781	0.4844	0.5001	0.9654	[0.0034702934336066026, 0.9653985402997675]	[0.003859968359802011, 0.9963822194552067]
0.4808	11.0	1232	0.3318	0.4845	0.5001	0.9665	[0.0024548211803361556, 0.9665399129138876]	[0.00263781478941615, 0.9975982819808147]
0.4808	12.0	1344	0.3552	0.4849	0.5005	0.9664	[0.0033778104561300974, 0.9663867278345344]	[0.003649486356013335, 0.9974086755418741]
0.4808	13.0	1456	0.2612	0.4845	0.5001	0.9672	[0.0017608302346806158, 0.9671985933973519]	[0.0018535995817518893, 0.9983025657191121]
0.3392	14.0	1568	0.2300	0.4845	0.5001	0.9671	[0.0018163185523506766, 0.9671249858066228]	[0.001916404695785607, 0.9982246340273064]
0.3392	15.0	1680	0.2657	0.4845	0.5001	0.9672	[0.0018194025597222916, 0.9671517415294609]	[0.001918102131300032, 0.9982521972361976]

Framework versions

• Transformers 4.37.0
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0