segformer-cloth-hoodie-30-steps

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

• Loss: 0.3849
• Mean Iou: 0.2926
• Mean Accuracy: 0.3657
• Overall Accuracy: 0.9048
• Accuracy Background: nan
• Accuracy Button: 0.0
• Accuracy Cuff: 0.1239
• Accuracy Drawstring: 0.1256
• Accuracy Hem: 0.2565
• Accuracy Hood: 0.9038
• Accuracy Pocket: 0.0
• Accuracy Sleeve: 0.9272
• Accuracy Torso: 0.9544
• Accuracy Zipper: 0.0
• Iou Background: 0.0
• Iou Button: 0.0
• Iou Cuff: 0.0975
• Iou Drawstring: 0.1042
• Iou Hem: 0.1177
• Iou Hood: 0.8538
• Iou Pocket: 0.0
• Iou Sleeve: 0.8862
• Iou Torso: 0.8661
• Iou Zipper: 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: 8
• eval_batch_size: 8
• seed: 1337
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: polynomial
• training_steps: 10000

Training results

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.2
• Datasets 2.18.0
• Tokenizers 0.15.2