segformer-b0-scene-parse-150-lr-3-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: 0.1523
• Mean Iou: 0.5014
• Mean Accuracy: 0.5220
• Overall Accuracy: 0.9615
• Per Category Iou: [0.04132646470292031, 0.9614038983247747]
• Per Category Accuracy: [0.053216300812732126, 0.9907305584765508]

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: 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	0.1629	0.4844	0.5	0.9688	[0.0, 0.9687870873345269]	[0.0, 1.0]
No log	2.0	224	0.1437	0.4844	0.5000	0.9688	[2.03629353850122e-05, 0.968778060560053]	[2.0369226173097684e-05, 0.9999900466190115]
No log	3.0	336	0.1551	0.4844	0.5	0.9688	[0.0, 0.9687870873345269]	[0.0, 1.0]
No log	4.0	448	0.1536	0.4873	0.5029	0.9674	[0.0072237010873418455, 0.967349403560223]	[0.0076096034111664095, 0.998278830733678]
0.254	5.0	560	0.1730	0.4844	0.5000	0.9688	[1.697363485298286e-06, 0.9687858141149847]	[1.697435514424807e-06, 0.9999986327773367]
0.254	6.0	672	0.1726	0.4844	0.5000	0.9688	[0.0, 0.9687868224249946]	[0.0, 0.9999997265554673]
0.254	7.0	784	0.1418	0.4886	0.5042	0.9679	[0.009270700532836455, 0.9678754695078028]	[0.009627854237817505, 0.998758780577388]
0.254	8.0	896	0.1618	0.4844	0.5	0.9688	[0.0, 0.9687870873345269]	[0.0, 1.0]
0.2012	9.0	1008	0.1350	0.4868	0.5023	0.9685	[0.005035086692148778, 0.9684816005292253]	[0.005109280898418669, 0.9995252456024103]
0.2012	10.0	1120	0.1429	0.4975	0.5137	0.9673	[0.027791805303191197, 0.967227089869692]	[0.02998689579782864, 0.997455270490238]
0.2012	11.0	1232	0.1419	0.4852	0.5008	0.9688	[0.0015964088435281823, 0.9688182225729328]	[0.0015972868190737434, 0.9999822807942842]
0.2012	12.0	1344	0.1339	0.4872	0.5028	0.9686	[0.00582435621561196, 0.968612834428971]	[0.00589010123505408, 0.9996363187715734]
0.2012	13.0	1456	0.1422	0.4990	0.5165	0.9652	[0.03289244256624029, 0.9651360857253766]	[0.03794447348945214, 0.9950514742926044]
0.1837	14.0	1568	0.1423	0.4928	0.5087	0.9673	[0.01828545458590366, 0.9672482875211772]	[0.019532390464486255, 0.9978029278690511]
0.1837	15.0	1680	0.1523	0.5014	0.5220	0.9615	[0.04132646470292031, 0.9614038983247747]	[0.053216300812732126, 0.9907305584765508]

Framework versions

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