detr-r101-cd45rb-8ah-12l

This model is a fine-tuned version of facebook/detr-resnet-101 on the cd45rb_nan_xywh dataset. It achieves the following results on the evaluation set:

• Loss: 1.6447

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: 1e-05
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 25
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
3.0401	1.0	4606	2.1637
2.7151	2.0	9212	2.0666
2.6021	3.0	13818	1.9868
2.5221	4.0	18424	1.8958
2.4541	5.0	23030	1.8810
2.4155	6.0	27636	1.8369
2.3531	7.0	32242	1.8040
2.31	8.0	36848	1.7979
2.2841	9.0	41454	1.7521
2.2555	10.0	46060	1.7243
2.3388	11.0	50666	1.8520
2.3523	12.0	55272	1.8499
2.3515	13.0	59878	1.7635
2.3236	14.0	64484	1.7787
2.2676	15.0	69090	1.7518
2.2787	16.0	73696	1.7879
2.2523	17.0	78302	1.7303
2.2357	18.0	82908	1.7361
2.2068	19.0	87514	1.6916
2.1972	20.0	92120	1.6941
2.1856	21.0	96726	1.6824
2.1611	22.0	101332	1.6711
2.1419	23.0	105938	1.6535
2.1412	24.0	110544	1.6602
2.1285	25.0	115150	1.6447

Framework versions

• Transformers 4.28.0
• Pytorch 2.0.1
• Datasets 2.12.0
• Tokenizers 0.13.3