detr-r50-cd45rb-8ah-6l-dilation-corrected

This model is a fine-tuned version of facebook/detr-resnet-50 on the cd45rb dataset. It achieves the following results on the evaluation set:

• Loss: 1.5202

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
2.4024	1.0	4606	1.7613
2.1798	2.0	9212	1.7139
2.1158	3.0	13818	1.6784
2.0907	4.0	18424	1.6514
2.0665	5.0	23030	1.6573
2.0511	6.0	27636	1.6508
2.0401	7.0	32242	1.6145
2.0217	8.0	36848	1.6353
2.0119	9.0	41454	1.6176
1.9921	10.0	46060	1.6012
1.9841	11.0	50666	1.5832
1.9774	12.0	55272	1.6204
1.9567	13.0	59878	1.5836
1.9542	14.0	64484	1.5789
1.9347	15.0	69090	1.5565
1.9348	16.0	73696	1.5833
1.9188	17.0	78302	1.5547
1.9085	18.0	82908	1.5456
1.8956	19.0	87514	1.5433
1.8891	20.0	92120	1.5555
1.8899	21.0	96726	1.5278
1.8782	22.0	101332	1.5235
1.8676	23.0	105938	1.5314
1.8699	24.0	110544	1.5172
1.8627	25.0	115150	1.5202

Framework versions

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