segformer-b0-finetuned-segments-gear2

This model is a fine-tuned version of nvidia/mit-b0 on the marcomameli01/gear dataset. It achieves the following results on the evaluation set:

• Loss: 0.1268
• Mean Iou: 0.1254
• Mean Accuracy: 0.2509
• Overall Accuracy: 0.2509
• Per Category Iou: [0.0, 0.2508641975308642]
• Per Category Accuracy: [nan, 0.2508641975308642]

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	Per Category Iou	Per Category Accuracy
0.4614	5.0	20	0.4427	0.0741	0.1481	0.1481	[0.0, 0.14814814814814814]	[nan, 0.14814814814814814]
0.3327	10.0	40	0.2933	0.1726	0.3453	0.3453	[0.0, 0.34528395061728395]	[nan, 0.34528395061728395]
0.2305	15.0	60	0.2244	0.0382	0.0763	0.0763	[0.0, 0.07634567901234568]	[nan, 0.07634567901234568]
0.2011	20.0	80	0.2130	0.0374	0.0748	0.0748	[0.0, 0.07476543209876543]	[nan, 0.07476543209876543]
0.1846	25.0	100	0.1672	0.1037	0.2073	0.2073	[0.0, 0.20730864197530866]	[nan, 0.20730864197530866]
0.1622	30.0	120	0.1532	0.0805	0.1611	0.1611	[0.0, 0.1610864197530864]	[nan, 0.1610864197530864]
0.139	35.0	140	0.1396	0.0971	0.1942	0.1942	[0.0, 0.19417283950617284]	[nan, 0.19417283950617284]
0.1342	40.0	160	0.1283	0.0748	0.1496	0.1496	[0.0, 0.14962962962962964]	[nan, 0.14962962962962964]
0.128	45.0	180	0.1224	0.1128	0.2256	0.2256	[0.0, 0.22558024691358025]	[nan, 0.22558024691358025]
0.1243	50.0	200	0.1268	0.1254	0.2509	0.2509	[0.0, 0.2508641975308642]	[nan, 0.2508641975308642]

Framework versions

• Transformers 4.20.0
• Pytorch 1.11.0+cu113
• Datasets 2.3.2
• Tokenizers 0.12.1