yq-msdetr-v1

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

• Loss: 0.2997
• Map: 0.8558
• Map 50: 0.9711
• Map 75: 0.958
• Map Small: -1.0
• Map Medium: 0.7923
• Map Large: 0.8688
• Mar 1: 0.1068
• Mar 10: 0.8028
• Mar 100: 0.8945
• Mar Small: -1.0
• Mar Medium: 0.8318
• Mar Large: 0.9076
• Map Per Class: -1.0
• Mar 100 Per Class: -1.0
• Classes: 0

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: 8.844509159584469e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Map	Map 50	Map 75	Map Small	Map Medium	Map Large	Mar 1	Mar 10	Mar 100	Mar Small	Mar Medium	Mar Large	Map Per Class	Mar 100 Per Class	Classes
1.1349	1.0	60	1.1033	0.108	0.1727	0.1227	-1.0	0.1053	0.1084	0.0607	0.1384	0.1384	-1.0	0.1352	0.139	-1.0	-1.0	0
0.539	2.0	120	0.6247	0.6349	0.8732	0.7464	-1.0	0.5186	0.6661	0.0847	0.6607	0.7696	-1.0	0.7381	0.7762	-1.0	-1.0	0
0.6188	3.0	180	0.4613	0.7527	0.9409	0.872	-1.0	0.6877	0.7662	0.0972	0.7303	0.831	-1.0	0.7871	0.8402	-1.0	-1.0	0
0.5357	4.0	240	0.5072	0.7236	0.8941	0.8423	-1.0	0.5907	0.754	0.0974	0.709	0.7989	-1.0	0.6496	0.8302	-1.0	-1.0	0
0.512	5.0	300	0.3948	0.7914	0.9496	0.9143	-1.0	0.7033	0.8124	0.0996	0.756	0.8457	-1.0	0.7524	0.8653	-1.0	-1.0	0
0.441	6.0	360	0.3634	0.8159	0.9523	0.9266	-1.0	0.7406	0.8317	0.1036	0.7737	0.8656	-1.0	0.7905	0.8813	-1.0	-1.0	0
0.4963	7.0	420	0.3878	0.805	0.9484	0.9145	-1.0	0.6978	0.8296	0.1019	0.7696	0.8573	-1.0	0.7456	0.8808	-1.0	-1.0	0
0.4507	8.0	480	0.3827	0.804	0.9531	0.9144	-1.0	0.7222	0.821	0.1036	0.7663	0.8551	-1.0	0.7705	0.8729	-1.0	-1.0	0
0.4929	9.0	540	0.3638	0.816	0.958	0.9208	-1.0	0.7508	0.8302	0.1047	0.7746	0.8698	-1.0	0.8192	0.8804	-1.0	-1.0	0
0.3256	10.0	600	0.3206	0.8441	0.9692	0.9459	-1.0	0.7803	0.8582	0.1057	0.7955	0.8858	-1.0	0.8292	0.8977	-1.0	-1.0	0
0.362	11.0	660	0.3044	0.8539	0.972	0.9484	-1.0	0.7985	0.8668	0.1054	0.8021	0.8923	-1.0	0.837	0.9038	-1.0	-1.0	0
0.3345	12.0	720	0.3051	0.8515	0.9708	0.947	-1.0	0.7886	0.8631	0.1059	0.8	0.8902	-1.0	0.8304	0.9027	-1.0	-1.0	0
0.3457	13.0	780	0.2982	0.8555	0.9715	0.9493	-1.0	0.7938	0.8694	0.106	0.8032	0.8947	-1.0	0.8307	0.9081	-1.0	-1.0	0
0.3684	14.0	840	0.2981	0.8573	0.9712	0.9582	-1.0	0.7967	0.8685	0.1072	0.804	0.895	-1.0	0.8355	0.9075	-1.0	-1.0	0
0.3411	15.0	900	0.2997	0.8558	0.9711	0.958	-1.0	0.7923	0.8688	0.1068	0.8028	0.8945	-1.0	0.8318	0.9076	-1.0	-1.0	0

Framework versions

• Transformers 4.41.1
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1