detr-r50-cd45rb-8ah-4l-1cost

This model is a fine-tuned version of facebook/detr-resnet-50 on the cd45rb_nan_xywh dataset. It achieves the following results on the evaluation set:

• Loss: 1.0859

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-06
• 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: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.7758	1.0	4606	1.3048
1.6314	2.0	9212	1.2281
1.5709	3.0	13818	1.1944
1.5368	4.0	18424	1.1702
1.5132	5.0	23030	1.1497
1.4929	6.0	27636	1.1419
1.4803	7.0	32242	1.1298
1.4703	8.0	36848	1.1246
1.4655	9.0	41454	1.1222
1.4625	10.0	46060	1.1222
1.4579	11.0	50666	1.1178
1.4529	12.0	55272	1.1086
1.4397	13.0	59878	1.1034
1.4366	14.0	64484	1.0990
1.425	15.0	69090	1.0943
1.4252	16.0	73696	1.0912
1.4215	17.0	78302	1.0881
1.4188	18.0	82908	1.0875
1.4133	19.0	87514	1.0870
1.4152	20.0	92120	1.0859

Framework versions

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