yq-fbdetr-v2

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: 0.2067
• Map: 0.9152
• Map 50: 0.9897
• Map 75: 0.979
• Map Small: -1.0
• Map Medium: 0.8667
• Map Large: 0.9242
• Mar 1: 0.1114
• Mar 10: 0.8451
• Mar 100: 0.939
• Mar Small: -1.0
• Mar Medium: 0.9
• Mar Large: 0.9472
• 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: 4.0400782557182323e-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: 29

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
0.5824	1.0	120	0.6091	0.6119	0.7625	0.7089	-1.0	0.566	0.6253	0.0967	0.6459	0.6837	-1.0	0.6418	0.6925	-1.0	-1.0	0
0.4222	2.0	240	0.4189	0.7908	0.9563	0.9095	-1.0	0.7371	0.8036	0.1015	0.7565	0.8548	-1.0	0.7971	0.8669	-1.0	-1.0	0
0.6961	3.0	360	0.3489	0.8233	0.9644	0.9283	-1.0	0.7686	0.8346	0.1022	0.7833	0.873	-1.0	0.8109	0.8861	-1.0	-1.0	0
0.4016	4.0	480	0.3397	0.8341	0.9692	0.9342	-1.0	0.7777	0.849	0.1046	0.7867	0.8772	-1.0	0.8195	0.8893	-1.0	-1.0	0
0.3081	5.0	600	0.3015	0.8575	0.9765	0.9647	-1.0	0.8129	0.8663	0.1064	0.8016	0.8916	-1.0	0.8501	0.9003	-1.0	-1.0	0
0.231	6.0	720	0.2907	0.8619	0.9754	0.9619	-1.0	0.81	0.874	0.1056	0.8099	0.9013	-1.0	0.8464	0.9128	-1.0	-1.0	0
0.3146	7.0	840	0.3052	0.8548	0.9728	0.9495	-1.0	0.7966	0.8682	0.1058	0.8024	0.8929	-1.0	0.8387	0.9042	-1.0	-1.0	0
0.2933	8.0	960	0.2768	0.8675	0.978	0.9667	-1.0	0.8173	0.8796	0.1074	0.8122	0.9019	-1.0	0.853	0.9121	-1.0	-1.0	0
0.2518	9.0	1080	0.2807	0.8683	0.9777	0.9668	-1.0	0.8184	0.8785	0.1076	0.8092	0.9001	-1.0	0.8562	0.9093	-1.0	-1.0	0
0.2744	10.0	1200	0.2709	0.8773	0.9785	0.9673	-1.0	0.8314	0.8868	0.1087	0.817	0.9087	-1.0	0.8653	0.9178	-1.0	-1.0	0
0.3472	11.0	1320	0.2631	0.8825	0.9791	0.9684	-1.0	0.8378	0.8907	0.1087	0.822	0.9118	-1.0	0.8682	0.9209	-1.0	-1.0	0
0.2634	12.0	1440	0.2761	0.8672	0.9878	0.9669	-1.0	0.8144	0.8787	0.1064	0.81	0.901	-1.0	0.8547	0.9107	-1.0	-1.0	0
0.2601	13.0	1560	0.2480	0.8857	0.9797	0.969	-1.0	0.8409	0.8953	0.1092	0.8248	0.9158	-1.0	0.8771	0.924	-1.0	-1.0	0
0.236	14.0	1680	0.2343	0.8971	0.9894	0.9791	-1.0	0.8432	0.9101	0.1086	0.8341	0.9247	-1.0	0.8736	0.9354	-1.0	-1.0	0
0.2099	15.0	1800	0.2367	0.8957	0.9892	0.9782	-1.0	0.8439	0.9057	0.1103	0.8332	0.9232	-1.0	0.8748	0.9334	-1.0	-1.0	0
0.2469	16.0	1920	0.2354	0.8975	0.9894	0.9785	-1.0	0.8489	0.9084	0.1099	0.8324	0.9244	-1.0	0.8782	0.9341	-1.0	-1.0	0
0.2666	17.0	2040	0.2367	0.893	0.9887	0.9776	-1.0	0.8473	0.9022	0.1093	0.8306	0.9224	-1.0	0.8788	0.9316	-1.0	-1.0	0
0.3254	18.0	2160	0.2290	0.902	0.9894	0.9786	-1.0	0.8494	0.9134	0.1099	0.8349	0.9279	-1.0	0.8811	0.9377	-1.0	-1.0	0
0.2248	19.0	2280	0.2206	0.9024	0.9894	0.9785	-1.0	0.8565	0.9151	0.1107	0.837	0.9302	-1.0	0.8871	0.9392	-1.0	-1.0	0
0.196	20.0	2400	0.2315	0.8986	0.9894	0.9785	-1.0	0.8549	0.9092	0.1107	0.835	0.9271	-1.0	0.8845	0.936	-1.0	-1.0	0
0.2661	21.0	2520	0.2173	0.9048	0.9897	0.9791	-1.0	0.8557	0.917	0.1107	0.8381	0.9315	-1.0	0.8885	0.9405	-1.0	-1.0	0
0.1741	22.0	2640	0.2182	0.9071	0.9897	0.9794	-1.0	0.8538	0.9185	0.111	0.8399	0.9326	-1.0	0.8871	0.9422	-1.0	-1.0	0
0.2364	23.0	2760	0.2172	0.9076	0.9896	0.979	-1.0	0.854	0.9175	0.1108	0.8384	0.9318	-1.0	0.8854	0.9415	-1.0	-1.0	0
0.2472	24.0	2880	0.2110	0.911	0.9896	0.9787	-1.0	0.8603	0.92	0.1106	0.8426	0.9362	-1.0	0.8937	0.9452	-1.0	-1.0	0
0.2159	25.0	3000	0.2130	0.9106	0.9896	0.9789	-1.0	0.8646	0.9191	0.1109	0.8415	0.9355	-1.0	0.8989	0.9432	-1.0	-1.0	0
0.2412	26.0	3120	0.2084	0.9142	0.9897	0.9788	-1.0	0.864	0.922	0.1113	0.8443	0.9382	-1.0	0.8954	0.9471	-1.0	-1.0	0
0.2208	27.0	3240	0.2101	0.9123	0.9897	0.9789	-1.0	0.8645	0.9211	0.1112	0.8425	0.9373	-1.0	0.8983	0.9455	-1.0	-1.0	0
0.1643	28.0	3360	0.2066	0.915	0.9895	0.9788	-1.0	0.8668	0.9238	0.1115	0.8446	0.9389	-1.0	0.9003	0.947	-1.0	-1.0	0
0.2868	29.0	3480	0.2067	0.9152	0.9897	0.979	-1.0	0.8667	0.9242	0.1114	0.8451	0.939	-1.0	0.9	0.9472	-1.0	-1.0	0

Framework versions

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