detr-resnet-50-dc5-fashionpedia-finetuned

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: 2.5675
• Map: 0.0275
• Map 50: 0.0634
• Map 75: 0.0182
• Map Small: 0.0
• Map Medium: 0.032
• Map Large: 0.0752
• Mar 1: 0.0195
• Mar 10: 0.0979
• Mar 100: 0.1196
• Mar Small: 0.0
• Mar Medium: 0.1307
• Mar Large: 0.3012
• Map Aortic enlargement: 0.1155
• Mar 100 Aortic enlargement: 0.4321
• Map Cardiomegaly: 0.132
• Mar 100 Cardiomegaly: 0.5943
• Map Ild: 0.0
• Mar 100 Ild: 0.0
• Map Lung opacity: 0.0001
• Mar 100 Lung opacity: 0.05
• Map Nodule/mass: 0.0
• Mar 100 Nodule/mass: 0.0
• Map Other lesion: 0.0
• Mar 100 Other lesion: 0.0
• Map Pleural effusion: 0.0
• Mar 100 Pleural effusion: 0.0
• Map Pleural thickening: 0.0
• Mar 100 Pleural thickening: 0.0
• Map Pulmonary fibrosis: 0.0
• Mar 100 Pulmonary fibrosis: 0.0

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: 1e-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: 1000
• 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	Map Aortic enlargement	Mar 100 Aortic enlargement	Map Atelectasis	Mar 100 Atelectasis	Map Cardiomegaly	Mar 100 Cardiomegaly	Map Ild	Mar 100 Ild	Map Infiltration	Mar 100 Infiltration	Map Lung opacity	Mar 100 Lung opacity	Map Nodule/mass	Mar 100 Nodule/mass	Map Other lesion	Mar 100 Other lesion	Map Pleural effusion	Mar 100 Pleural effusion	Map Pleural thickening	Mar 100 Pleural thickening	Map Pulmonary fibrosis	Mar 100 Pulmonary fibrosis
No log	2.7778	50	3.9076	0.0006	0.0028	0.0002	0.0	0.0003	0.005	0.0	0.0057	0.021	0.0	0.0143	0.0845	0.0006	0.0415	-1.0	-1.0	0.0047	0.1472	0.0	0.0	-1.0	-1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.3075	5.5556	100	3.5001	0.0047	0.0127	0.0028	0.0	0.0033	0.0271	0.0	0.0317	0.0849	0.0	0.0774	0.2357	0.0223	0.3377	-1.0	-1.0	0.0199	0.4264	0.0	0.0	-1.0	-1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.3075	8.3333	150	3.2511	0.0079	0.0197	0.0044	0.0	0.005	0.0459	0.0029	0.0465	0.099	0.0	0.0962	0.3083	0.0296	0.3057	0.0411	0.5849	0.0	0.0	-1.0	-1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
3.7002	11.1111	200	3.0549	0.0084	0.0247	0.0047	0.0	0.0073	0.0297	0.0055	0.053	0.096	0.0	0.1022	0.2488	0.0317	0.3811	0.0439	0.483	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
3.7002	13.8889	250	2.8268	0.0121	0.0276	0.0078	0.0	0.0073	0.052	0.0101	0.0591	0.1115	0.0	0.1193	0.3083	0.0356	0.4	0.0733	0.6038	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
3.3402	16.6667	300	2.8859	0.016	0.0383	0.0123	0.0	0.0303	0.0413	0.0073	0.0857	0.1061	0.0	0.1212	0.2476	0.0591	0.4604	0.0846	0.4943	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
3.3402	19.4444	350	2.8537	0.0166	0.0363	0.0124	0.0	0.0342	0.0553	0.0084	0.0748	0.1	0.0	0.1113	0.2393	0.0381	0.4264	0.1117	0.4736	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
3.2237	22.2222	400	2.7335	0.0177	0.0419	0.0131	0.0	0.0277	0.0508	0.0096	0.0925	0.1107	0.0	0.1143	0.2833	0.0629	0.4509	0.0963	0.5453	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
3.2237	25.0	450	2.7128	0.0209	0.0499	0.0155	0.0	0.0236	0.0642	0.0157	0.0922	0.1145	0.0	0.1176	0.2964	0.0741	0.4679	0.114	0.5623	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
3.0321	27.7778	500	2.6574	0.0197	0.0433	0.0137	0.0	0.035	0.0498	0.0143	0.0973	0.1143	0.0	0.1192	0.2881	0.082	0.4717	0.095	0.5566	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
3.0321	30.5556	550	2.6041	0.0206	0.0504	0.0158	0.0	0.0375	0.0616	0.0184	0.0939	0.1202	0.0	0.1204	0.3036	0.0693	0.4585	0.1157	0.5736	0.0	0.0	0.0001	0.05	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
2.9455	33.3333	600	2.6164	0.0252	0.0548	0.0214	0.0	0.0413	0.0575	0.0212	0.0979	0.1174	0.0	0.1286	0.3083	0.1141	0.4472	0.1131	0.6094	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
2.9455	36.1111	650	2.5827	0.023	0.0516	0.0172	0.0	0.0321	0.0486	0.0191	0.0948	0.1212	0.0	0.1322	0.3119	0.1166	0.4755	0.0908	0.6151	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
2.8715	38.8889	700	2.5590	0.028	0.0587	0.0232	0.0	0.0333	0.0607	0.0247	0.0992	0.1413	0.0	0.1453	0.3143	0.1387	0.4717	0.1123	0.6	0.0	0.0	0.0006	0.2	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
2.8715	41.6667	750	2.5453	0.0264	0.0579	0.0192	0.0	0.0273	0.0678	0.0243	0.0975	0.1193	0.0	0.1275	0.3155	0.1166	0.4566	0.121	0.617	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
2.7836	44.4444	800	2.5490	0.0249	0.0559	0.0173	0.0	0.0291	0.0634	0.0226	0.0992	0.1166	0.0	0.1223	0.3083	0.1078	0.4509	0.1163	0.5981	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
2.7836	47.2222	850	2.5600	0.0256	0.0593	0.0187	0.0	0.034	0.0628	0.0189	0.0962	0.1195	0.0	0.1292	0.3083	0.1153	0.4698	0.1152	0.6057	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
2.7591	50.0	900	2.5640	0.0258	0.0574	0.0185	0.0	0.0336	0.0636	0.0205	0.1039	0.1219	0.0	0.1319	0.3036	0.1159	0.4509	0.1162	0.5962	0.0	0.0	0.0001	0.05	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
2.7591	52.7778	950	2.5807	0.0272	0.0621	0.0175	0.0	0.0308	0.0777	0.0203	0.0981	0.1149	0.0	0.1252	0.3036	0.109	0.4377	0.1361	0.5962	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
2.7396	55.5556	1000	2.5675	0.0275	0.0634	0.0182	0.0	0.032	0.0752	0.0195	0.0979	0.1196	0.0	0.1307	0.3012	0.1155	0.4321	0.132	0.5943	0.0	0.0	0.0001	0.05	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0

Framework versions

• Transformers 4.46.2
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3