segformer-b0-scene-parse-150

This model is a fine-tuned version of nvidia/mit-b0 on the scene_parse_150 dataset. It achieves the following results on the evaluation set:

• Loss: 2.4363

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: 25

Training results

Training Loss	Epoch	Step	Validation Loss
4.9247	1.0	20	4.9059
4.5746	2.0	40	4.6125
4.5749	3.0	60	4.0543
4.2089	4.0	80	3.7518
3.9694	5.0	100	3.5456
3.4601	6.0	120	3.2747
3.5999	7.0	140	3.2017
3.2724	8.0	160	2.9125
3.3775	9.0	180	2.9575
2.9466	10.0	200	2.7341
2.8341	11.0	220	2.7660
3.0457	12.0	240	2.8243
3.8739	13.0	260	2.6063
2.6287	14.0	280	2.6507
2.6291	15.0	300	2.5783
3.1508	16.0	320	2.4557
2.7062	17.0	340	2.3839
3.0388	18.0	360	2.5159
2.1882	19.0	380	2.4833
2.4467	20.0	400	2.4029
2.2945	21.0	420	2.4240
2.5678	22.0	440	2.4354
2.8432	23.0	460	2.4213
2.6837	24.0	480	2.3866
2.398	25.0	500	2.4363

Framework versions

• Transformers 4.44.2
• Pytorch 1.11.0
• Datasets 2.21.0
• Tokenizers 0.19.1