rfdetr-docvqa-media3-trainval-agree2-medium

This model is a fine-tuned version of Roboflow/rf-detr-medium on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 15.6134
• Map: 0.283
• Map 50: 0.444
• Map 75: 0.2928
• Map Small: -1.0
• Map Medium: 0.0168
• Map Large: 0.3019
• Mar 1: 0.2934
• Mar 10: 0.5636
• Mar 100: 0.6339
• Mar Small: -1.0
• Mar Medium: 0.0417
• Mar Large: 0.6672
• Map Chart: 0.3127
• Mar 100 Chart: 0.7731
• Map Image: 0.348
• Mar 100 Image: 0.6877
• Map Signature: 0.1883
• Mar 100 Signature: 0.4408

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: 16
• 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 Chart	Mar 100 Chart	Map Image	Mar 100 Image	Map Signature	Mar 100 Signature
108.3258	1.0	27	29.7238	0.0001	0.0006	0.0	-1.0	0.0	0.0001	0.0012	0.0076	0.0328	-1.0	0.0	0.036	0.0	0.0154	0.0002	0.0789	0.0	0.0041
33.7814	2.0	54	24.9789	0.0003	0.0009	0.0	-1.0	0.0	0.0003	0.0018	0.0053	0.014	-1.0	0.0	0.0151	0.0	0.0038	0.0007	0.0158	0.0	0.0224
16.5911	3.0	81	24.3980	0.0005	0.0019	0.0002	-1.0	0.0	0.0005	0.0023	0.0209	0.0598	-1.0	0.0	0.0644	0.0004	0.0538	0.0011	0.1053	0.0001	0.0204
14.5205	4.0	108	18.7300	0.0013	0.004	0.0004	-1.0	0.0	0.0014	0.0159	0.0519	0.1827	-1.0	0.0	0.2006	0.0	0.0	0.0017	0.3421	0.0023	0.2061
12.1968	5.0	135	19.3451	0.004	0.0095	0.002	-1.0	0.0	0.0043	0.0306	0.1037	0.2538	-1.0	0.0	0.2795	0.0	0.0	0.0077	0.5246	0.0043	0.2367
11.8570	6.0	162	18.1114	0.0048	0.0115	0.0029	-1.0	0.0038	0.0052	0.0318	0.1482	0.2916	-1.0	0.05	0.3173	0.0	0.0	0.0093	0.6035	0.005	0.2714
11.4029	7.0	189	17.6193	0.0182	0.0366	0.0132	-1.0	0.0092	0.0197	0.0739	0.2067	0.3533	-1.0	0.275	0.366	0.0013	0.1	0.0382	0.6333	0.0152	0.3265
10.9795	8.0	216	16.1149	0.0385	0.0774	0.0356	-1.0	0.0194	0.0414	0.1114	0.335	0.521	-1.0	0.1917	0.5419	0.0085	0.4808	0.075	0.7088	0.0321	0.3735
11.0632	9.0	243	14.9344	0.0437	0.0718	0.0425	-1.0	0.0486	0.0467	0.1281	0.3684	0.5406	-1.0	0.1417	0.5649	0.006	0.5462	0.086	0.686	0.039	0.3898
10.1947	10.0	270	14.7902	0.069	0.1093	0.0702	-1.0	0.121	0.0734	0.1889	0.424	0.6135	-1.0	0.175	0.6373	0.0114	0.6923	0.127	0.7175	0.0686	0.4306
10.4935	11.0	297	14.6825	0.0653	0.1041	0.069	-1.0	0.009	0.07	0.2103	0.4371	0.6194	-1.0	0.0917	0.6495	0.0135	0.7231	0.1074	0.7228	0.075	0.4122
10.0782	12.0	324	14.9505	0.0758	0.1363	0.0728	-1.0	0.0044	0.0811	0.1967	0.4021	0.6321	-1.0	0.1083	0.6604	0.0375	0.7731	0.1237	0.707	0.0662	0.4163
9.8978	13.0	351	14.3118	0.0699	0.1246	0.0636	-1.0	0.0489	0.0748	0.2423	0.4511	0.6483	-1.0	0.1917	0.6717	0.0206	0.7692	0.108	0.7368	0.081	0.4388
9.5010	14.0	378	14.9869	0.0874	0.14	0.0922	-1.0	0.0133	0.094	0.2449	0.5036	0.6457	-1.0	0.1	0.6753	0.0297	0.8038	0.1378	0.7333	0.0947	0.4
9.4850	15.0	405	14.1153	0.1425	0.2483	0.1337	-1.0	0.0285	0.1506	0.2616	0.5246	0.6717	-1.0	0.1583	0.6985	0.131	0.8154	0.1643	0.7526	0.1321	0.4469
9.1458	16.0	432	16.1527	0.1153	0.1919	0.1131	-1.0	0.0013	0.1256	0.2596	0.4627	0.6149	-1.0	0.0167	0.6497	0.0465	0.7423	0.2003	0.6982	0.0992	0.4041
9.0574	17.0	459	14.0260	0.1643	0.2785	0.1726	-1.0	0.0252	0.1736	0.2902	0.5473	0.6675	-1.0	0.1	0.699	0.194	0.7923	0.158	0.7509	0.1409	0.4592
9.0914	18.0	486	15.2129	0.2144	0.3435	0.2318	-1.0	0.0108	0.2268	0.2991	0.5649	0.6741	-1.0	0.05	0.709	0.2405	0.8231	0.2183	0.7298	0.1843	0.4694
8.9114	19.0	513	14.6411	0.221	0.3576	0.2225	-1.0	0.004	0.2347	0.291	0.5638	0.6647	-1.0	0.0333	0.7006	0.2415	0.8115	0.2195	0.7316	0.2021	0.451
8.6744	20.0	540	14.7526	0.2301	0.3709	0.23	-1.0	0.0515	0.2429	0.2909	0.5667	0.6833	-1.0	0.1083	0.715	0.3189	0.8	0.2462	0.7439	0.1251	0.5061
8.4979	21.0	567	15.3168	0.2433	0.3792	0.26	-1.0	0.003	0.2604	0.2843	0.5635	0.6511	-1.0	0.0333	0.6865	0.2627	0.7885	0.2949	0.7281	0.1722	0.4367
8.3844	22.0	594	15.0938	0.2543	0.4056	0.2731	-1.0	0.0185	0.2704	0.2941	0.5667	0.6597	-1.0	0.0583	0.6933	0.2765	0.8	0.3025	0.7158	0.184	0.4633
8.1649	23.0	621	14.9846	0.2254	0.3773	0.2192	-1.0	0.0311	0.2399	0.2838	0.5812	0.6454	-1.0	0.0833	0.6766	0.2499	0.7731	0.2536	0.714	0.1728	0.449
8.1756	24.0	648	15.3787	0.289	0.4327	0.2849	-1.0	0.0224	0.3068	0.3058	0.5927	0.6389	-1.0	0.05	0.6721	0.3623	0.7692	0.3251	0.6965	0.1797	0.451
8.1260	25.0	675	15.3256	0.2641	0.4384	0.2609	-1.0	0.009	0.2831	0.3098	0.5615	0.6467	-1.0	0.0333	0.6815	0.2984	0.7808	0.3159	0.6982	0.1779	0.4612
7.8612	26.0	702	15.3995	0.2981	0.4714	0.2976	-1.0	0.0196	0.3156	0.3017	0.5537	0.6373	-1.0	0.0417	0.6707	0.3596	0.7769	0.3493	0.686	0.1853	0.449
7.9190	27.0	729	15.5125	0.278	0.4423	0.2902	-1.0	0.0168	0.2951	0.3096	0.5595	0.6464	-1.0	0.0417	0.6804	0.3192	0.7808	0.3286	0.6912	0.1862	0.4673
7.8385	28.0	756	15.5406	0.2943	0.4484	0.3169	-1.0	0.0224	0.3119	0.3027	0.5606	0.6345	-1.0	0.05	0.6673	0.3636	0.7692	0.3413	0.6895	0.178	0.4449
7.9839	29.0	783	15.5843	0.2728	0.4273	0.2701	-1.0	0.014	0.2914	0.3025	0.5607	0.6368	-1.0	0.0333	0.6708	0.3053	0.7846	0.3382	0.6912	0.1749	0.4347
8.1108	30.0	810	15.6134	0.283	0.444	0.2928	-1.0	0.0168	0.3019	0.2934	0.5636	0.6339	-1.0	0.0417	0.6672	0.3127	0.7731	0.348	0.6877	0.1883	0.4408

Framework versions

• Transformers 5.12.1
• Pytorch 2.12.0+cu130
• Datasets 5.0.0
• Tokenizers 0.22.2