rtdetrv2-floorplan

This model is a fine-tuned version of PekingU/rtdetr_v2_r50vd on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 6.9335
• Map: 0.5424
• Map 50: 0.6551
• Map 75: 0.5889
• Map Small: 0.5058
• Map Medium: 0.7206
• Map Large: 0.2238
• Mar 1: 0.269
• Mar 10: 0.8104
• Mar 100: 0.862
• Mar Small: 0.8045
• Mar Medium: 0.8731
• Mar Large: 0.7548
• Map Window: 0.6578
• Mar 100 Window: 0.818
• Map Door: 0.4366
• Mar 100 Door: 0.9306
• Map Stair: 0.5329
• Mar 100 Stair: 0.8374

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: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 0.05
• num_epochs: 30

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 Window	Mar 100 Window	Map Door	Mar 100 Door	Map Stair	Mar 100 Stair
15.4458	1.0	578	10.1131	0.3164	0.471	0.3402	0.2361	0.4257	0.0229	0.197	0.6397	0.7017	0.5295	0.7788	0.7281	0.2784	0.6279	0.3905	0.8119	0.2801	0.6654
12.2832	2.0	1156	7.6855	0.1317	0.188	0.1413	0.3031	0.4809	0.0089	0.1955	0.637	0.7434	0.6543	0.7858	0.6864	0.0968	0.6455	0.1156	0.8598	0.1827	0.7249
11.4532	3.0	1734	7.5403	0.2899	0.4262	0.31	0.3112	0.5237	0.0306	0.2058	0.6804	0.7715	0.6889	0.8139	0.7395	0.3363	0.682	0.2229	0.8802	0.3104	0.7524
11.0336	4.0	2312	7.1088	0.4099	0.5728	0.4578	0.3675	0.6124	0.1748	0.2313	0.711	0.7861	0.7113	0.8385	0.7003	0.4868	0.7122	0.3246	0.8899	0.4182	0.7562
10.8294	5.0	2890	7.0414	0.289	0.427	0.2975	0.3075	0.6326	0.0113	0.2144	0.6826	0.7737	0.6865	0.8515	0.6741	0.2862	0.6613	0.3105	0.8909	0.2702	0.7689
9.9642	6.0	3468	6.9391	0.4075	0.5434	0.4544	0.3786	0.6975	0.0397	0.2429	0.7441	0.8055	0.7304	0.8647	0.8171	0.3638	0.7284	0.3945	0.8963	0.4641	0.7917
10.3160	7.0	4046	6.9321	0.4568	0.6163	0.5041	0.4145	0.6558	0.0723	0.2428	0.7325	0.7855	0.7077	0.8303	0.8912	0.4497	0.6784	0.4062	0.8991	0.5143	0.7789
9.6594	8.0	4624	6.9460	0.4711	0.6106	0.5182	0.4316	0.6786	0.1844	0.2578	0.7663	0.8291	0.7515	0.8716	0.7216	0.6212	0.7649	0.3505	0.9078	0.4416	0.8145
9.3666	9.0	5202	6.9855	0.4525	0.5849	0.5362	0.4849	0.6653	0.1748	0.2497	0.7501	0.8175	0.7502	0.8465	0.8968	0.5892	0.7311	0.2725	0.912	0.4958	0.8095
9.6941	10.0	5780	6.7884	0.4715	0.6195	0.5388	0.4316	0.6738	0.0865	0.2468	0.7576	0.8205	0.7497	0.8428	0.8961	0.5902	0.7374	0.3863	0.9122	0.4382	0.8118
9.2232	11.0	6358	7.1479	0.4814	0.6233	0.5488	0.4009	0.6939	0.2222	0.2627	0.7607	0.8242	0.7534	0.8563	0.7628	0.5623	0.7387	0.4037	0.9138	0.4782	0.8199
8.6241	12.0	6936	6.8049	0.5369	0.6939	0.6068	0.4684	0.6814	0.2824	0.2642	0.7789	0.8224	0.7561	0.845	0.7086	0.6019	0.7324	0.4551	0.9182	0.5539	0.8166
8.8883	13.0	7514	7.1613	0.3402	0.4516	0.3873	0.4198	0.5423	0.0242	0.2306	0.7415	0.8141	0.7396	0.8615	0.7725	0.4525	0.7198	0.2057	0.9181	0.3624	0.8043
8.5386	14.0	8092	6.9806	0.4852	0.622	0.5323	0.4613	0.6726	0.2121	0.2538	0.7773	0.8311	0.7661	0.8546	0.7407	0.6047	0.755	0.3629	0.92	0.4881	0.8183
8.4977	15.0	8670	6.7821	0.5355	0.6572	0.5973	0.4674	0.7286	0.2562	0.2768	0.8079	0.844	0.7836	0.8603	0.7534	0.5899	0.7788	0.5027	0.9253	0.5139	0.8279
8.5491	16.0	9248	6.9631	0.4955	0.621	0.5539	0.4554	0.7096	0.2082	0.2617	0.7867	0.8388	0.7788	0.8586	0.7346	0.619	0.764	0.4016	0.9234	0.4661	0.8289
8.4493	17.0	9826	6.9747	0.5573	0.6877	0.6192	0.4832	0.7432	0.2102	0.2743	0.8046	0.8498	0.7871	0.8728	0.7892	0.661	0.7806	0.4625	0.928	0.5485	0.8407
8.5492	18.0	10404	6.9642	0.5356	0.6735	0.5853	0.4805	0.6884	0.18	0.2646	0.8005	0.8409	0.7807	0.8689	0.7403	0.6078	0.764	0.4791	0.9268	0.52	0.8318
8.3569	19.0	10982	7.0092	0.5331	0.6578	0.584	0.4883	0.7133	0.238	0.2661	0.8021	0.8481	0.79	0.868	0.7255	0.6381	0.7829	0.434	0.9312	0.5273	0.8301
8.2804	20.0	11560	7.0399	0.5724	0.6933	0.6307	0.5121	0.7474	0.1573	0.2723	0.8074	0.8507	0.7946	0.8704	0.7407	0.6391	0.7923	0.5071	0.9295	0.571	0.8301
7.9573	21.0	12138	6.8740	0.4999	0.615	0.554	0.5034	0.706	0.1803	0.2648	0.7942	0.8511	0.7989	0.8647	0.758	0.6366	0.7941	0.3695	0.9302	0.4935	0.8289
8.1709	22.0	12716	6.9873	0.3753	0.4571	0.414	0.4797	0.6558	0.1876	0.2431	0.7734	0.8532	0.7949	0.8727	0.7414	0.4913	0.7874	0.3423	0.9282	0.2922	0.8441
8.3106	23.0	13294	6.9599	0.5574	0.6772	0.6062	0.4997	0.7208	0.2467	0.2719	0.8177	0.862	0.8044	0.8708	0.7929	0.6602	0.8126	0.4801	0.9309	0.5319	0.8426
7.7713	24.0	13872	6.9604	0.5582	0.6719	0.6061	0.4987	0.7173	0.2562	0.2744	0.813	0.8613	0.8049	0.8766	0.7488	0.6586	0.8126	0.473	0.9302	0.5431	0.8412
8.1088	25.0	14450	6.9547	0.5377	0.6443	0.584	0.4936	0.7283	0.2262	0.272	0.8079	0.8592	0.8045	0.8688	0.7531	0.6409	0.8059	0.45	0.9314	0.5221	0.8403
7.8254	26.0	15028	6.9371	0.5335	0.6346	0.58	0.4976	0.7273	0.2183	0.2699	0.8035	0.8569	0.801	0.8721	0.7424	0.6315	0.809	0.4431	0.928	0.5261	0.8337
7.7139	27.0	15606	6.9369	0.5769	0.6932	0.6298	0.5109	0.7263	0.2694	0.2721	0.8138	0.8614	0.8074	0.8735	0.7826	0.666	0.8144	0.5028	0.9293	0.5619	0.8405
7.9592	28.0	16184	6.9072	0.5373	0.6488	0.5817	0.5038	0.7235	0.2138	0.2638	0.8037	0.8609	0.8026	0.8696	0.7827	0.6494	0.8149	0.4506	0.9302	0.5119	0.8375
7.9454	29.0	16762	6.9385	0.565	0.6775	0.6137	0.5085	0.7275	0.2386	0.2725	0.8123	0.8615	0.8037	0.8729	0.7832	0.6605	0.8153	0.4839	0.9297	0.5506	0.8394
7.6347	30.0	17340	6.9335	0.5424	0.6551	0.5889	0.5058	0.7206	0.2238	0.269	0.8104	0.862	0.8045	0.8731	0.7548	0.6578	0.818	0.4366	0.9306	0.5329	0.8374

Framework versions

• Transformers 5.12.1
• Pytorch 2.12.1+cu130
• Datasets 5.0.0
• Tokenizers 0.22.2