polejowska/detr-r101-official

detr-r101-cd45rb-8ah-6l-256d

This model is a fine-tuned version of facebook/detr-resnet-101 on the cd45rb dataset. It achieves the following results on the evaluation set:

• Loss: 1.5675

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
2.4837	1.0	4606	1.7975
2.2753	2.0	9212	1.7412
2.2035	3.0	13818	1.7140
2.185	4.0	18424	1.7262
2.15	5.0	23030	1.6788
2.1109	6.0	27636	1.6751
2.0949	7.0	32242	1.6688
2.071	8.0	36848	1.6277
2.046	9.0	41454	1.6120
2.0338	10.0	46060	1.6056
2.0472	11.0	50666	1.6164
2.0389	12.0	55272	1.6055
2.0215	13.0	59878	1.5925
2.0123	14.0	64484	1.5913
1.9957	15.0	69090	1.5839
2.0014	16.0	73696	1.5774
2.0155	17.0	78302	1.5963
2.0142	18.0	82908	1.6062
2.0083	19.0	87514	1.5912
1.9949	20.0	92120	1.5875
1.9976	21.0	96726	1.5909
1.9816	22.0	101332	1.5811
1.9779	23.0	105938	1.5700
1.9706	24.0	110544	1.5680
1.96	25.0	115150	1.5675

Framework versions

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