detr-resnet-50_finetuned_bbatch

This model is a fine-tuned version of facebook/detr-resnet-50 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.3729

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 200
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
4.0202	4.17	50	3.2912
3.2043	8.33	100	3.1033
3.0326	12.5	150	3.0882
2.957	16.67	200	2.9938
2.8396	20.83	250	2.8404
2.7485	25.0	300	2.8785
2.6515	29.17	350	2.7908
2.5651	33.33	400	2.7162
2.5172	37.5	450	2.6793
2.434	41.67	500	2.6200
2.3921	45.83	550	2.6078
2.37	50.0	600	2.5634
2.3211	54.17	650	2.5730
2.2658	58.33	700	2.4883
2.2487	62.5	750	2.6039
2.2001	66.67	800	2.4645
2.1767	70.83	850	2.5015
2.1533	75.0	900	2.4727
2.1489	79.17	950	2.4663
2.1181	83.33	1000	2.4865
2.0739	87.5	1050	2.4522
2.0664	91.67	1100	2.4344
2.0347	95.83	1150	2.4347
2.0175	100.0	1200	2.4883
1.9989	104.17	1250	2.4523
1.9773	108.33	1300	2.4971
1.9473	112.5	1350	2.5026
1.9431	116.67	1400	2.4029
1.9274	120.83	1450	2.4887
1.907	125.0	1500	2.4216
1.9006	129.17	1550	2.4143
1.8775	133.33	1600	2.4107
1.8795	137.5	1650	2.3722
1.8582	141.67	1700	2.3952
1.8538	145.83	1750	2.3964
1.8409	150.0	1800	2.4083
1.8302	154.17	1850	2.4013
1.8203	158.33	1900	2.3768
1.8112	162.5	1950	2.4538
1.7935	166.67	2000	2.3903
1.8035	170.83	2050	2.3707
1.7955	175.0	2100	2.3588
1.7786	179.17	2150	2.3775
1.7988	183.33	2200	2.3376
1.7808	187.5	2250	2.3629
1.7802	191.67	2300	2.4191
1.769	195.83	2350	2.3897
1.7677	200.0	2400	2.3729

Framework versions

• Transformers 4.28.1
• Pytorch 2.0.0+cu118
• Datasets 2.11.0
• Tokenizers 0.13.3