marco_model

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: 0.3925
• Map: 0.7574
• Map 50: 0.929
• Map 75: 0.884
• Map Small: -1.0
• Map Medium: 0.6668
• Map Large: 0.7845
• Mar 1: 0.1077
• Mar 10: 0.7532
• Mar 100: 0.8489
• Mar Small: -1.0
• Mar Medium: 0.8074
• Mar Large: 0.8615
• 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: 5e-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: cosine
• num_epochs: 60

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
1.8485	1.0	9	2.0519	0.0	0.0	0.0	-1.0	0.0	0.0	0.0	0.0	0.0	-1.0	0.0	0.0	-1.0	-1.0	0
1.4965	2.0	18	1.4544	0.0	0.0	0.0	-1.0	0.0	0.0	0.0	0.0	0.0	-1.0	0.0	0.0	-1.0	-1.0	0
1.3577	3.0	27	1.3521	0.0	0.0	0.0	-1.0	0.0	0.0	0.0	0.0	0.0	-1.0	0.0	0.0	-1.0	-1.0	0
1.2729	4.0	36	1.2400	0.0077	0.0281	0.004	-1.0	0.0158	0.0058	0.0039	0.009	0.009	-1.0	0.0148	0.0073	-1.0	-1.0	0
1.1613	5.0	45	1.2042	0.0066	0.0099	0.0099	-1.0	0.0079	0.0059	0.006	0.006	0.006	-1.0	0.0148	0.0034	-1.0	-1.0	0
1.1631	6.0	54	1.1257	0.0235	0.0297	0.0229	-1.0	0.0139	0.0274	0.0219	0.0219	0.0219	-1.0	0.0167	0.0235	-1.0	-1.0	0
1.1196	7.0	63	1.1354	0.0231	0.0297	0.0297	-1.0	0.031	0.0234	0.0172	0.021	0.021	-1.0	0.0315	0.0179	-1.0	-1.0	0
1.123	8.0	72	1.0910	0.023	0.0297	0.0297	-1.0	0.0314	0.0168	0.0142	0.0176	0.0176	-1.0	0.0315	0.0134	-1.0	-1.0	0
1.0342	9.0	81	1.0235	0.0576	0.0748	0.0748	-1.0	0.0601	0.0618	0.0421	0.0622	0.0622	-1.0	0.0593	0.0631	-1.0	-1.0	0
1.0549	10.0	90	1.1559	0.0366	0.0495	0.0495	-1.0	0.0609	0.0297	0.0155	0.0352	0.0352	-1.0	0.0593	0.0279	-1.0	-1.0	0
0.9125	11.0	99	0.9385	0.1832	0.2508	0.2184	-1.0	0.2598	0.1648	0.0691	0.2189	0.2253	-1.0	0.3111	0.1994	-1.0	-1.0	0
0.8747	12.0	108	0.9022	0.3659	0.5151	0.4412	-1.0	0.3361	0.394	0.0824	0.4519	0.5056	-1.0	0.5481	0.4927	-1.0	-1.0	0
0.7734	13.0	117	0.7992	0.3737	0.5225	0.4491	-1.0	0.3209	0.3989	0.0828	0.4468	0.4605	-1.0	0.4444	0.4654	-1.0	-1.0	0
0.9409	14.0	126	0.7235	0.474	0.6527	0.5714	-1.0	0.3893	0.5149	0.0974	0.5451	0.5798	-1.0	0.5259	0.5961	-1.0	-1.0	0
0.6904	15.0	135	0.6478	0.5892	0.7933	0.7054	-1.0	0.4855	0.6278	0.0948	0.6292	0.7236	-1.0	0.6778	0.7374	-1.0	-1.0	0
0.6725	16.0	144	0.5890	0.6553	0.8521	0.7901	-1.0	0.5227	0.7068	0.0953	0.6734	0.7725	-1.0	0.687	0.7983	-1.0	-1.0	0
0.6808	17.0	153	0.5885	0.6474	0.8739	0.7781	-1.0	0.5653	0.6778	0.0948	0.6476	0.7867	-1.0	0.7537	0.7966	-1.0	-1.0	0
0.4056	18.0	162	0.5272	0.6694	0.8891	0.8089	-1.0	0.5952	0.697	0.0944	0.679	0.8013	-1.0	0.7741	0.8095	-1.0	-1.0	0
0.5985	19.0	171	0.5146	0.655	0.8835	0.7829	-1.0	0.5655	0.6914	0.0966	0.6738	0.7914	-1.0	0.7537	0.8028	-1.0	-1.0	0
0.395	20.0	180	0.5012	0.6867	0.8954	0.8308	-1.0	0.5623	0.7287	0.0987	0.6742	0.806	-1.0	0.7537	0.8218	-1.0	-1.0	0
0.5581	21.0	189	0.4747	0.7079	0.9051	0.8362	-1.0	0.5806	0.7501	0.094	0.7	0.8227	-1.0	0.7926	0.8318	-1.0	-1.0	0
0.3999	22.0	198	0.4860	0.7029	0.9115	0.8551	-1.0	0.6319	0.7286	0.1004	0.6948	0.812	-1.0	0.7815	0.8212	-1.0	-1.0	0
0.4602	23.0	207	0.4910	0.6976	0.8884	0.8222	-1.0	0.6346	0.7229	0.1017	0.6983	0.8009	-1.0	0.7889	0.8045	-1.0	-1.0	0
0.4216	24.0	216	0.4817	0.7081	0.902	0.8552	-1.0	0.5966	0.7483	0.1	0.709	0.8176	-1.0	0.7796	0.8291	-1.0	-1.0	0
0.4984	25.0	225	0.4645	0.7113	0.9192	0.8414	-1.0	0.6434	0.7322	0.1056	0.7155	0.8004	-1.0	0.7519	0.8151	-1.0	-1.0	0
0.6411	26.0	234	0.4574	0.712	0.9219	0.8573	-1.0	0.6225	0.7431	0.1013	0.706	0.8189	-1.0	0.7944	0.8263	-1.0	-1.0	0
0.4605	27.0	243	0.4536	0.7152	0.9188	0.8555	-1.0	0.6354	0.7416	0.103	0.7099	0.8219	-1.0	0.8019	0.8279	-1.0	-1.0	0
0.4985	28.0	252	0.4349	0.7105	0.9045	0.8509	-1.0	0.6036	0.7465	0.1017	0.721	0.8159	-1.0	0.7407	0.8385	-1.0	-1.0	0
0.3637	29.0	261	0.4239	0.7218	0.9085	0.8706	-1.0	0.5929	0.7671	0.1043	0.7343	0.8176	-1.0	0.7185	0.8475	-1.0	-1.0	0
0.3068	30.0	270	0.4442	0.7149	0.9153	0.8518	-1.0	0.6273	0.7437	0.1034	0.7202	0.8227	-1.0	0.7926	0.8318	-1.0	-1.0	0
0.3456	31.0	279	0.4668	0.7161	0.9101	0.8587	-1.0	0.6638	0.7374	0.1039	0.7189	0.8129	-1.0	0.7722	0.8251	-1.0	-1.0	0
0.437	32.0	288	0.4049	0.7589	0.9395	0.8984	-1.0	0.6825	0.7851	0.1034	0.7391	0.8455	-1.0	0.8037	0.8581	-1.0	-1.0	0
0.3444	33.0	297	0.4229	0.747	0.9414	0.8875	-1.0	0.6868	0.7668	0.1043	0.7266	0.8288	-1.0	0.7889	0.8408	-1.0	-1.0	0
0.3523	34.0	306	0.4203	0.7474	0.9398	0.8744	-1.0	0.6821	0.7687	0.1077	0.7343	0.8305	-1.0	0.7963	0.8408	-1.0	-1.0	0
0.2629	35.0	315	0.4216	0.7449	0.928	0.8813	-1.0	0.6819	0.7653	0.1094	0.7361	0.8313	-1.0	0.7778	0.8475	-1.0	-1.0	0
0.3533	36.0	324	0.4698	0.7009	0.8817	0.8321	-1.0	0.638	0.7228	0.0974	0.709	0.8185	-1.0	0.7926	0.8263	-1.0	-1.0	0
0.2833	37.0	333	0.4804	0.7025	0.8904	0.8546	-1.0	0.6607	0.718	0.0996	0.6983	0.8137	-1.0	0.7944	0.8196	-1.0	-1.0	0
0.3656	38.0	342	0.4546	0.7221	0.9065	0.8543	-1.0	0.6382	0.7488	0.1039	0.727	0.8193	-1.0	0.7593	0.8374	-1.0	-1.0	0
0.4347	39.0	351	0.4096	0.7362	0.905	0.8617	-1.0	0.6388	0.7679	0.112	0.7365	0.8343	-1.0	0.7667	0.8547	-1.0	-1.0	0
0.3386	40.0	360	0.4118	0.7517	0.9218	0.8883	-1.0	0.6665	0.777	0.1021	0.7416	0.8468	-1.0	0.8019	0.8603	-1.0	-1.0	0
0.2718	41.0	369	0.4035	0.7532	0.9305	0.8954	-1.0	0.6756	0.7727	0.109	0.7425	0.8429	-1.0	0.8093	0.8531	-1.0	-1.0	0
0.2767	42.0	378	0.4029	0.7598	0.9324	0.8973	-1.0	0.6522	0.7911	0.1112	0.7442	0.8446	-1.0	0.7796	0.8642	-1.0	-1.0	0
0.3407	43.0	387	0.4037	0.7574	0.9348	0.8848	-1.0	0.6612	0.7855	0.1086	0.7416	0.8433	-1.0	0.7852	0.8609	-1.0	-1.0	0
0.2914	44.0	396	0.3938	0.7646	0.9385	0.9041	-1.0	0.6736	0.7913	0.109	0.7455	0.8442	-1.0	0.787	0.8615	-1.0	-1.0	0
0.2816	45.0	405	0.3961	0.7667	0.9377	0.8954	-1.0	0.671	0.794	0.1064	0.7519	0.8476	-1.0	0.7963	0.8631	-1.0	-1.0	0
0.2724	46.0	414	0.4053	0.754	0.9263	0.8842	-1.0	0.6651	0.7787	0.106	0.7472	0.8378	-1.0	0.7778	0.8559	-1.0	-1.0	0
0.2173	47.0	423	0.4031	0.7515	0.9222	0.879	-1.0	0.653	0.7808	0.106	0.7442	0.8395	-1.0	0.7815	0.857	-1.0	-1.0	0
0.3299	48.0	432	0.4073	0.7563	0.9291	0.8775	-1.0	0.662	0.7839	0.1069	0.7498	0.8442	-1.0	0.7963	0.8587	-1.0	-1.0	0
0.326	49.0	441	0.3962	0.7601	0.929	0.8834	-1.0	0.6728	0.7868	0.1073	0.7515	0.8515	-1.0	0.8093	0.8642	-1.0	-1.0	0
0.267	50.0	450	0.3949	0.7592	0.9284	0.8917	-1.0	0.6737	0.7834	0.1064	0.7506	0.8506	-1.0	0.8111	0.8626	-1.0	-1.0	0
0.3073	51.0	459	0.3979	0.754	0.9278	0.8833	-1.0	0.6652	0.7805	0.1064	0.7451	0.8451	-1.0	0.8056	0.857	-1.0	-1.0	0
0.3166	52.0	468	0.3995	0.7545	0.9279	0.8822	-1.0	0.6645	0.78	0.106	0.7446	0.8446	-1.0	0.8074	0.8559	-1.0	-1.0	0
0.2635	53.0	477	0.3957	0.7554	0.9279	0.8826	-1.0	0.6605	0.7836	0.1073	0.7468	0.8464	-1.0	0.8037	0.8592	-1.0	-1.0	0
0.2663	54.0	486	0.3942	0.7556	0.9273	0.8819	-1.0	0.6625	0.7831	0.1073	0.7524	0.8476	-1.0	0.8037	0.8609	-1.0	-1.0	0
0.3247	55.0	495	0.3940	0.7574	0.9272	0.8819	-1.0	0.6688	0.7832	0.1073	0.7536	0.8494	-1.0	0.8093	0.8615	-1.0	-1.0	0
0.2399	56.0	504	0.3933	0.7575	0.9275	0.8825	-1.0	0.6671	0.7836	0.1073	0.7532	0.8494	-1.0	0.8093	0.8615	-1.0	-1.0	0
0.3087	57.0	513	0.3929	0.7591	0.929	0.8839	-1.0	0.6682	0.7855	0.1077	0.7536	0.8498	-1.0	0.8093	0.862	-1.0	-1.0	0
0.2753	58.0	522	0.3926	0.7575	0.929	0.884	-1.0	0.6667	0.7843	0.1077	0.7532	0.8489	-1.0	0.8074	0.8615	-1.0	-1.0	0
0.2291	59.0	531	0.3925	0.7574	0.929	0.8839	-1.0	0.6667	0.7845	0.1077	0.7532	0.8489	-1.0	0.8074	0.8615	-1.0	-1.0	0
0.2809	60.0	540	0.3925	0.7574	0.929	0.884	-1.0	0.6668	0.7845	0.1077	0.7532	0.8489	-1.0	0.8074	0.8615	-1.0	-1.0	0

Framework versions

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