detr-r101-cd45rb-8ah-4l

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: 1.7442

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
3.1054	1.0	4606	2.2520
2.7633	2.0	9212	2.1251
2.6589	3.0	13818	2.0489
2.5832	4.0	18424	2.0000
2.5369	5.0	23030	1.9442
2.4955	6.0	27636	1.9363
2.4615	7.0	32242	1.8860
2.4326	8.0	36848	1.8732
2.404	9.0	41454	1.8580
2.3915	10.0	46060	1.8536
2.4311	11.0	50666	1.8881
2.4175	12.0	55272	1.8443
2.3763	13.0	59878	1.8380
2.371	14.0	64484	1.8314
2.3427	15.0	69090	1.8255
2.3415	16.0	73696	1.8258
2.3189	17.0	78302	1.7938
2.3049	18.0	82908	1.7959
2.2884	19.0	87514	1.7827
2.2769	20.0	92120	1.7849
2.2728	21.0	96726	1.7581
2.2547	22.0	101332	1.7553
2.249	23.0	105938	1.7521
2.2476	24.0	110544	1.7505
2.2371	25.0	115150	1.7442

Framework versions

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