inspection_finetuned

This model is a fine-tuned version of facebook/detr-resnet-50 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.4719
• Map: 0.0122
• Map 50: 0.0264
• Map 75: 0.011
• Map Small: 0.0024
• Map Medium: 0.0143
• Map Large: 0.0207
• Mar 1: 0.1017
• Mar 10: 0.1856
• Mar 100: 0.2427
• Mar Small: 0.0494
• Mar Medium: 0.2108
• Mar Large: 0.3076
• Map Cut: 0.0186
• Mar 100 Cut: 0.5419
• Map Hole: 0.0
• Mar 100 Hole: 0.0
• Map Stain: 0.0303
• Mar 100 Stain: 0.4289
• Map Threaderror: 0.0
• Mar 100 Threaderror: 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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss	Map	Map 50	Map 75	Map Small	Map Medium	Map Large	Mar 1	Mar 10	Mar 100	Mar Small	Mar Medium	Mar Large	Map Cut	Mar 100 Cut	Map Hole	Mar 100 Hole	Map Stain	Mar 100 Stain	Map Threaderror	Mar 100 Threaderror
No log	1.0	167	1.5593	0.0093	0.0179	0.0094	0.0018	0.0125	0.0194	0.09	0.1791	0.2239	0.0411	0.1774	0.2914	0.0198	0.5158	0.0	0.0	0.0174	0.3797	0.0	0.0
No log	2.0	334	1.4719	0.0122	0.0264	0.011	0.0024	0.0143	0.0207	0.1017	0.1856	0.2427	0.0494	0.2108	0.3076	0.0186	0.5419	0.0	0.0	0.0303	0.4289	0.0	0.0

Framework versions

• Transformers 4.52.4
• Pytorch 2.6.0+cu124
• Datasets 2.14.4
• Tokenizers 0.21.1