rtdetr-v2-r50-cardamage-40ep

This model is a fine-tuned version of PekingU/rtdetr_r50vd on an unknown dataset. It achieves the following results on the evaluation set:

• eval_loss: 12.7731
• eval_model_preparation_time: 0.0097
• eval_map: 0.2093
• eval_map_50: 0.404
• eval_map_75: 0.1925
• eval_map_small: 0.0168
• eval_map_medium: 0.0853
• eval_map_large: 0.2916
• eval_mar_1: 0.3526
• eval_mar_10: 0.4898
• eval_mar_100: 0.5314
• eval_mar_small: 0.0167
• eval_mar_medium: 0.2585
• eval_mar_large: 0.6342
• eval_map_car-parts: -1.0
• eval_mar_100_car-parts: -1.0
• eval_map_Bonet: 0.1129
• eval_mar_100_Bonet: 0.3364
• eval_map_Bumper: 0.128
• eval_mar_100_Bumper: 0.4396
• eval_map_Door: 0.1729
• eval_mar_100_Door: 0.6609
• eval_map_Headlight: 0.1512
• eval_mar_100_Headlight: 0.4947
• eval_map_Mirror: 0.2436
• eval_mar_100_Mirror: 0.5529
• eval_map_Tailight: 0.345
• eval_mar_100_Tailight: 0.575
• eval_map_Windshield: 0.3114
• eval_mar_100_Windshield: 0.66
• eval_runtime: 10.9627
• eval_samples_per_second: 14.048
• eval_steps_per_second: 1.824
• step: 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 adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 300
• num_epochs: 40
• mixed_precision_training: Native AMP

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0