ft-segformer-with-sceneparse150

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: 4.3884
• Mean Iou: 0.0259
• Mean Accuracy: 0.0547
• Overall Accuracy: 0.4491
• Per Category Iou: [0.3417101480816873, 0.4211288769068327, 0.7747555282866107, 0.3846204354053868, 0.33732378973954696, 0.041151599293209766, 0.46128131427542346, 0.11439788718514722, 0.12616558604979503, 0.18171159576156137, 0.17165912703264458, 0.06346386631243024, 0.11546430134541383, 0.0001487343415790393, 0.0013247427763715854, 0.0, 0.13274620610379087, 0.004944101773323053, 0.011655503401719319, 0.0, 0.0016660546838606434, 0.0, 0.035477393149597074, 0.0, 0.0, 4.4454718423813505e-06, 0.06028847248426353, 0.0, 0.0006802721088435374, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0005368298173512513, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0012449941795398464, 0.0, 0.0, 0.0, 0.0, 0.0, nan, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.00042048608191068876, 0.0, 0.012187069195213215, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.00016971877598818757, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, nan, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 2.889338341519792e-05, 0.0, 0.0, 0.0, 0.0, 0.0029921675613834376, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.00030254393296857133, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, nan, 0.0]
• Per Category Accuracy: [0.5194495194945754, 0.8117698640339073, 0.9672763096625787, 0.820432246643049, 0.6843419269871048, 0.04300891893063284, 0.6009645810887155, 0.16730232665390735, 0.5003207343883315, 0.19801025930029267, 0.35768755152514425, 0.09390059524438853, 0.1317835995063082, 0.00014920000378920644, 0.0015635305528612998, 0.0, 0.2614850183183669, 0.009323204419889503, 0.04031575979701156, 0.0, 0.0017305272984988814, 0.0, 0.11935812364496419, 0.0, 0.0, 5.206923124986983e-06, 0.1372276664160497, 0.0, 0.0007093682075720508, 0.0, 0.0, nan, 0.0, 0.0, 0.001442127818781674, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0012546312652299554, 0.0, 0.0, nan, 0.0, 0.0, nan, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, nan, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0005804672761573066, 0.0, 0.03457228301948799, 0.0, 0.0, 0.0, nan, 0.0, 0.0, 0.00036832412523020257, nan, 0.0, nan, 0.0, nan, 0.0, 0.0, 0.0, 0.0, 0.0, nan, 0.0, 0.0, 0.0, nan, 0.0, 0.0, 0.0, nan, 0.0, 0.0, 0.0, nan, 0.0, 0.0, 0.0, 0.0, nan, 0.0, nan, nan, 0.0, 0.0, nan, nan, 0.0, 6.113964294448521e-05, nan, 0.0, nan, nan, 0.006153289295086417, nan, nan, 0.0, 0.0, 0.0, nan, 0.0, 0.0, 0.0, 0.0, 0.0, nan, nan, 0.0, nan, 0.0017814547540686615, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, nan, 0.0, nan, nan, nan, 0.0, 0.0, nan, 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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Framework versions

• Transformers 4.35.2
• Pytorch 2.0.0
• Datasets 2.15.0
• Tokenizers 0.15.0