MalyO2/working

facebook/detr-resnet-50-dc5

This model is a fine-tuned version of facebook/detr-resnet-50-dc5 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5836
• Map: 0.5257
• Map 50: 0.6508
• Map 75: 0.6241
• Map Small: 0.0
• Map Medium: 0.4752
• Map Large: 0.7513
• Mar 1: 0.1853
• Mar 10: 0.6
• Mar 100: 0.7147
• Mar Small: 0.0
• Mar Medium: 0.6684
• Mar Large: 0.8923

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: 5e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• training_steps: 400
• mixed_precision_training: Native AMP

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
4.1002	0.7692	10	4.1741	0.0003	0.001	0.0003	0.0	0.0062	0.0002	0.0	0.0	0.0441	0.0	0.0474	0.0462
1.772	1.5385	20	1.4577	0.0298	0.05	0.0286	0.0	0.0185	0.0656	0.0294	0.1206	0.4882	0.0	0.3421	0.7769
1.5665	2.3077	30	1.3869	0.0339	0.0549	0.0351	0.0	0.0407	0.0516	0.0029	0.0824	0.6059	0.0	0.5158	0.8308
2.0258	3.0769	40	1.2246	0.0561	0.0797	0.0593	0.0	0.0398	0.1166	0.0265	0.1206	0.6441	0.0	0.5789	0.8385
1.5082	3.8462	50	1.1988	0.0477	0.0869	0.0542	0.0	0.0927	0.063	0.0235	0.0853	0.6471	0.0	0.6316	0.7692
1.3716	4.6154	60	1.1917	0.0549	0.1014	0.0602	0.0	0.0902	0.0761	0.0588	0.1618	0.5971	0.0	0.5421	0.7692
1.2398	5.3846	70	1.0554	0.1329	0.1674	0.1485	0.0	0.1462	0.1957	0.0765	0.1882	0.7294	0.0	0.7474	0.8154
1.401	6.1538	80	0.9179	0.1176	0.1821	0.1315	0.0	0.0835	0.2295	0.0529	0.1794	0.7294	0.0	0.7211	0.8538
2.0328	6.9231	90	0.9198	0.1361	0.2109	0.1554	0.0	0.0937	0.2424	0.0559	0.2088	0.6882	0.0	0.6368	0.8692
1.6358	7.6923	100	0.9298	0.2252	0.2898	0.2523	0.0	0.2279	0.3487	0.1059	0.3176	0.6882	0.0	0.6263	0.8846
0.8849	8.4615	110	0.8894	0.1893	0.2435	0.2248	0.0	0.1438	0.3337	0.0971	0.2265	0.7265	0.0	0.7263	0.8385
1.1906	9.2308	120	0.8505	0.2105	0.2704	0.2598	0.0	0.1879	0.3317	0.1324	0.2706	0.6853	0.0	0.6474	0.8462
1.0404	10.0	130	0.7320	0.2508	0.2998	0.29	0.0	0.2031	0.4149	0.1588	0.2971	0.7471	0.0	0.7421	0.8692
1.1534	10.7692	140	0.7996	0.2832	0.374	0.3479	0.0	0.2502	0.411	0.1676	0.3647	0.6647	0.0	0.6263	0.8231
1.1725	11.5385	150	0.7990	0.3115	0.4464	0.3745	0.0	0.2972	0.4147	0.1294	0.3735	0.6588	0.0	0.6158	0.8231
0.891	12.3077	160	0.9007	0.2856	0.3519	0.3449	0.0	0.2607	0.3788	0.1029	0.3529	0.6735	0.0	0.6263	0.8462
1.1	13.0769	170	0.7376	0.2642	0.3608	0.3377	0.0	0.2281	0.4018	0.1176	0.3676	0.7176	0.0	0.7	0.8538
1.2631	13.8462	180	0.7162	0.306	0.4363	0.3899	0.0	0.2997	0.3933	0.1412	0.45	0.7059	0.0	0.7053	0.8154
1.0496	14.6154	190	0.7276	0.2811	0.3866	0.3483	0.0	0.3061	0.3685	0.1471	0.3882	0.7235	0.0	0.7316	0.8231
0.8883	15.3846	200	0.6855	0.3373	0.4578	0.4385	0.0	0.3441	0.4654	0.15	0.4824	0.7412	0.0	0.7579	0.8308
0.8471	16.1538	210	0.6733	0.4351	0.5932	0.5367	0.0	0.3702	0.6215	0.15	0.5412	0.7206	0.0	0.7158	0.8385
0.9084	16.9231	220	0.6526	0.4279	0.5632	0.4848	0.0	0.4011	0.572	0.1824	0.5647	0.7294	0.0	0.7105	0.8692
0.8872	17.6923	230	0.6218	0.4376	0.5753	0.5274	0.0	0.3879	0.6215	0.1559	0.5853	0.7382	0.0	0.7263	0.8692
0.9739	18.4615	240	0.6590	0.4494	0.6293	0.505	0.0	0.3889	0.65	0.1471	0.5853	0.7029	0.0	0.6895	0.8308
0.7596	19.2308	250	0.6367	0.4625	0.6229	0.5322	0.0	0.4106	0.6581	0.1529	0.5853	0.7118	0.0	0.7053	0.8308
0.7124	20.0	260	0.6601	0.4619	0.6411	0.5327	0.0	0.39	0.6852	0.1559	0.5765	0.6794	0.0	0.6421	0.8385
0.8369	20.7692	270	0.6363	0.4736	0.64	0.5738	0.0	0.3993	0.737	0.1559	0.5853	0.6853	0.0	0.6474	0.8462
0.8608	21.5385	280	0.6304	0.496	0.6406	0.5583	0.0	0.4484	0.6973	0.1588	0.5912	0.7	0.0	0.6579	0.8692
0.6174	22.3077	290	0.6825	0.4808	0.6714	0.5569	0.0	0.4264	0.6738	0.1529	0.5765	0.6735	0.0	0.6158	0.8615
0.5903	23.0769	300	0.6037	0.5187	0.6804	0.6126	0.0	0.4604	0.709	0.1824	0.6118	0.7206	0.0	0.6842	0.8846
0.6325	23.8462	310	0.6373	0.529	0.6819	0.6246	0.0	0.4489	0.7601	0.1765	0.5941	0.7088	0.0	0.6579	0.8923
0.8569	24.6154	320	0.6131	0.5382	0.6684	0.6357	0.0	0.4862	0.7382	0.1794	0.6147	0.7294	0.0	0.7	0.8846
0.7056	25.3846	330	0.5700	0.5244	0.6545	0.6089	0.0	0.4891	0.6871	0.1824	0.6176	0.75	0.0	0.7421	0.8769
0.5988	26.1538	340	0.5738	0.5437	0.7119	0.651	0.0	0.5362	0.6823	0.1853	0.6206	0.7529	0.0	0.7579	0.8615
0.5209	26.9231	350	0.6136	0.5153	0.6944	0.6047	0.0	0.4772	0.7054	0.1824	0.5882	0.7059	0.0	0.6789	0.8538
0.6547	27.6923	360	0.6338	0.5166	0.6645	0.6224	0.0	0.4842	0.7072	0.1882	0.5971	0.7088	0.0	0.6842	0.8538
0.6324	28.4615	370	0.6083	0.5143	0.6543	0.6279	0.0	0.4683	0.729	0.1853	0.6	0.7118	0.0	0.6789	0.8692
0.6323	29.2308	380	0.5748	0.529	0.6552	0.637	0.0	0.48	0.7529	0.1853	0.6088	0.7206	0.0	0.6842	0.8846
0.4509	30.0	390	0.5758	0.5311	0.652	0.6325	0.0	0.4923	0.7454	0.1882	0.6206	0.7324	0.0	0.7053	0.8846
0.8259	30.7692	400	0.5836	0.5257	0.6508	0.6241	0.0	0.4752	0.7513	0.1853	0.6	0.7147	0.0	0.6684	0.8923

Framework versions

• Transformers 4.46.3
• Pytorch 2.4.0
• Datasets 3.1.0
• Tokenizers 0.20.0