ModelListedExternal

Dfine Cppe5

This model is a fine-tuned version of ustc-community/dfine-small-coco on the cppe-5 dataset. It achieves the following results on the evaluation set: Loss: 4.3105 Map: 0.1854 Map 50: 0.2891 Map 75: 0.2037 Map Small:...

Sourcehuggingface.co/nonl/dfine-cppe5/tree/482f1df52cf0150b9f46db0baa91b613d3f765d2

KindOtherBase modelustc-community/dfine-small-cocoVersionRevision 482f1df

Model source

Kind: Other
File size: 39.2 MB
Base model: ustc-community/dfine-small-coco
Version: vRevision 482f1df
Updated: Sep 27, 2026
Downloads: 28
Tasks: Object detection
Source: Hugging Face

Source description

dfine-cppe5

This model is a fine-tuned version of ustc-community/dfine-small-coco on the cppe-5 dataset. It achieves the following results on the evaluation set:

Loss: 4.3105
Map: 0.1854
Map 50: 0.2891
Map 75: 0.2037
Map Small: 0.0615
Map Medium: 0.1352
Map Large: 0.2007
Mar 1: 0.2026
Mar 10: 0.4969
Mar 100: 0.5489
Mar Small: 0.3591
Mar Medium: 0.4387
Mar Large: 0.6796
Map Coverall: 0.6348
Mar 100 Coverall: 0.7464
Map Face Shield: 0.0897
Mar 100 Face Shield: 0.5885
Map Gloves: 0.0228
Mar 100 Gloves: 0.3697
Map Goggles: 0.0629
Mar 100 Goggles: 0.5239
Map Mask: 0.1167
Mar 100 Mask: 0.516

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 adamw_torch_fused with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 300.0

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 Coverall	Mar 100 Coverall	Map Face Shield	Mar 100 Face Shield	Map Gloves	Mar 100 Gloves	Map Goggles	Mar 100 Goggles	Map Mask	Mar 100 Mask
No log	1.0	107	4.7397	0.1272	0.2199	0.1199	0.0238	0.0845	0.1525	0.1638	0.389	0.4963	0.3193	0.3995	0.6218	0.4858	0.744	0.0094	0.4607	0.0629	0.4057	0.0036	0.35	0.0743	0.5211
No log	2.0	214	3.9616	0.241	0.4137	0.2407	0.1166	0.1831	0.3048	0.2865	0.5315	0.6025	0.4928	0.5156	0.706	0.5714	0.7881	0.0931	0.6033	0.1405	0.4474	0.1212	0.5891	0.279	0.5845
No log	3.0	321	3.9319	0.2392	0.3918	0.2347	0.0988	0.1515	0.3213	0.3109	0.5318	0.5995	0.4796	0.4894	0.7118	0.6084	0.7929	0.1166	0.6164	0.0936	0.4657	0.1715	0.5935	0.2059	0.5289
No log	4.0	428	3.9498	0.2396	0.3815	0.2458	0.0651	0.1566	0.328	0.2979	0.5336	0.5957	0.4184	0.5073	0.709	0.602	0.8071	0.1565	0.6098	0.0744	0.4657	0.1459	0.5652	0.2192	0.5304
21.0412	5.0	535	3.8719	0.2654	0.4173	0.2761	0.1002	0.1622	0.3524	0.3085	0.5551	0.6104	0.4359	0.5035	0.7216	0.6378	0.8048	0.2018	0.6607	0.0458	0.464	0.1993	0.5717	0.242	0.551
21.0412	6.0	642	3.8112	0.2551	0.3955	0.2703	0.0947	0.1752	0.333	0.2942	0.5441	0.6127	0.4041	0.5136	0.7249	0.6422	0.8095	0.2108	0.6459	0.0566	0.468	0.145	0.5761	0.2207	0.5639
21.0412	7.0	749	3.8236	0.2568	0.4011	0.2777	0.0846	0.1902	0.3321	0.2836	0.5435	0.6156	0.4832	0.5138	0.7206	0.6514	0.8065	0.2567	0.6525	0.0421	0.4794	0.1376	0.5761	0.1961	0.5634
21.0412	8.0	856	3.8126	0.2431	0.3778	0.2599	0.0726	0.1656	0.3299	0.2893	0.5415	0.6208	0.4274	0.52	0.73	0.6479	0.7958	0.2078	0.618	0.0511	0.4943	0.1539	0.6283	0.155	0.5675
21.0412	9.0	963	3.7309	0.2714	0.4186	0.296	0.0863	0.1662	0.3554	0.3179	0.5599	0.6357	0.4967	0.5527	0.7346	0.6597	0.803	0.2451	0.6721	0.0685	0.5006	0.1742	0.6239	0.2096	0.5789
16.0885	10.0	1070	3.7597	0.2574	0.4025	0.2786	0.0787	0.1702	0.3416	0.3014	0.55	0.626	0.4208	0.5356	0.725	0.6523	0.7964	0.244	0.6656	0.0747	0.4966	0.1211	0.5978	0.1948	0.5737
16.0885	11.0	1177	3.7166	0.2677	0.4142	0.2884	0.0736	0.1833	0.3557	0.311	0.5545	0.6332	0.4929	0.5321	0.7363	0.6608	0.7899	0.2733	0.6738	0.0511	0.5057	0.1345	0.6087	0.219	0.5881
16.0885	12.0	1284	3.6186	0.2689	0.4157	0.2939	0.0653	0.1893	0.3639	0.3269	0.5569	0.6336	0.4436	0.5529	0.7323	0.6605	0.7839	0.2282	0.6852	0.0786	0.5103	0.1541	0.6022	0.2229	0.5866
16.0885	13.0	1391	3.6877	0.2601	0.402	0.2831	0.0917	0.1805	0.3363	0.2964	0.5644	0.6341	0.4545	0.5233	0.7416	0.6674	0.7988	0.229	0.6393	0.0593	0.5194	0.1187	0.6261	0.2262	0.5866
16.0885	14.0	1498	3.7621	0.2439	0.3748	0.2631	0.0666	0.1607	0.3222	0.2838	0.5485	0.6248	0.3906	0.5133	0.734	0.6639	0.7946	0.2481	0.6361	0.0399	0.5057	0.103	0.6022	0.1647	0.5856
15.1531	15.0	1605	3.7746	0.235	0.3683	0.2535	0.062	0.1502	0.3003	0.2804	0.5472	0.6201	0.3918	0.5218	0.7258	0.6611	0.7845	0.2299	0.6541	0.0324	0.4897	0.0721	0.5848	0.1795	0.5876
15.1531	16.0	1712	3.7271	0.2402	0.3761	0.2476	0.068	0.1883	0.301	0.2945	0.554	0.6313	0.4119	0.5338	0.7332	0.6614	0.7988	0.1743	0.6492	0.0735	0.5206	0.0804	0.6087	0.2115	0.5794
15.1531	17.0	1819	3.6995	0.2622	0.4072	0.2828	0.0745	0.1621	0.3404	0.2987	0.5624	0.6411	0.4169	0.533	0.7517	0.6715	0.797	0.2643	0.6852	0.0419	0.4971	0.1061	0.6304	0.227	0.5959
15.1531	18.0	1926	3.7130	0.2486	0.3872	0.2598	0.0622	0.1503	0.3142	0.2854	0.5452	0.6323	0.355	0.5296	0.7553	0.6648	0.806	0.2647	0.6623	0.049	0.5269	0.0856	0.587	0.1792	0.5794
14.5821	19.0	2033	3.7003	0.2644	0.4098	0.282	0.0722	0.2056	0.3418	0.2978	0.5573	0.6422	0.4042	0.5388	0.7495	0.6608	0.7976	0.2702	0.6721	0.0503	0.5211	0.1202	0.6348	0.2204	0.5856
14.5821	20.0	2140	3.6850	0.2622	0.407	0.2831	0.0693	0.1767	0.3415	0.313	0.5565	0.645	0.4006	0.5436	0.7582	0.6668	0.7929	0.2388	0.6885	0.0451	0.5309	0.0796	0.6152	0.2808	0.5974
14.5821	21.0	2247	3.7363	0.2463	0.3775	0.2635	0.0557	0.1603	0.3236	0.2745	0.551	0.6365	0.4838	0.54	0.746	0.6675	0.7988	0.2276	0.6902	0.0393	0.5126	0.0661	0.587	0.2309	0.5938
14.5821	22.0	2354	3.8106	0.2383	0.3691	0.2527	0.0777	0.1528	0.2936	0.2696	0.5401	0.6172	0.3829	0.5142	0.7317	0.6642	0.7958	0.1825	0.6475	0.0389	0.4834	0.0652	0.5717	0.2408	0.5876
14.5821	23.0	2461	3.7385	0.2515	0.3986	0.2628	0.0663	0.1658	0.3241	0.3014	0.5564	0.6409	0.4272	0.5585	0.7422	0.6513	0.7899	0.2374	0.6902	0.049	0.524	0.0816	0.6087	0.238	0.5918
14.1787	24.0	2568	3.7549	0.2444	0.3787	0.2592	0.0689	0.161	0.3082	0.2842	0.5598	0.6359	0.483	0.5299	0.7504	0.6587	0.7881	0.2427	0.6934	0.0407	0.5166	0.0728	0.5804	0.2073	0.601
14.1787	25.0	2675	3.6648	0.2654	0.4129	0.2758	0.0692	0.1744	0.3496	0.3018	0.5649	0.6228	0.3994	0.527	0.7381	0.6672	0.7869	0.2299	0.6623	0.0782	0.4903	0.0753	0.5696	0.2763	0.6052
14.1787	26.0	2782	3.7563	0.2554	0.4015	0.2717	0.086	0.1806	0.3276	0.3037	0.5619	0.6423	0.4954	0.5506	0.745	0.6639	0.7863	0.2338	0.7098	0.0458	0.5246	0.0827	0.5957	0.2507	0.5954
14.1787	27.0	2889	3.7496	0.2614	0.4102	0.2823	0.0788	0.1558	0.3382	0.2991	0.5649	0.6254	0.4087	0.5288	0.7306	0.6544	0.7815	0.2311	0.6738	0.0633	0.4897	0.061	0.5826	0.2973	0.5995
14.1787	28.0	2996	3.7521	0.2464	0.3859	0.2587	0.0956	0.1416	0.3077	0.278	0.5592	0.6287	0.3855	0.5288	0.74	0.6656	0.7887	0.1633	0.659	0.0621	0.5069	0.0718	0.5935	0.2692	0.5954
13.8439	29.0	3103	3.7792	0.2442	...

Sources

1 source

Hugging FacePrimaryVersion 482f1df52cf0150b9f46db0baa91b613d3f765d2

huggingface.co/nonl/dfine-cppe5/tree/482f1df52cf0150b9f46db0baa91b613d3f765d2

Verified Sep 27

Model artifacts

1 artifact

model.safetensors

safetensors · 39.2 MB · SHA-256 5360724c24ab…9534 · Hugging Face

Download

Source excerpts

2 excerpts

ReadmeSep 27huggingface.co/nonl/dfine-cppe5

--- library_name: transformers license: apache-2.0 base_model: ustc-community/dfine-small-coco tags: - object-detection - vision - generated_from_trainer model-index: - name: dfine-cppe5 results: [] ---  # dfine-cppe5 This model is a fine-tuned version of [ustc-community/dfine-small-coco](https://huggingface.co/ustc-community/dfine-small-coco) on the cppe-5 dataset. It achieves the following results on the evaluation set: - Loss: 4.3105 - Map: 0.1854 - Map 50: 0.2891 - Map 75: 0.2037 - Map Small: 0.0615 - Map Medium: 0.1352 - Map Large: 0.2007 - Mar 1: 0.2026 - Mar 10: 0.4969 - Mar 100: 0.5489 - Mar Small: 0.3591 - Mar Medium: 0.4387 - Mar Large: 0.6796 - Map Coverall: 0.6348 - Mar 100 Coverall: 0.7464 - Map Face Shield: 0.0897 - Mar 100 Face Shield: 0.5885 - Map Gloves: 0.0228 - Mar 100 Gloves: 0.3697 - Map Goggles: 0.0629 - Mar 100 Goggles: 0.5239 - Map Mask: 0.1167 - Mar 100 Mask: 0.516 ## 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 adamw_torch_fused with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: linear - num_epochs: 300.0 ### 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 Coverall | Mar 100 Coverall | Map Face Shield | Mar 100 Face Shield | Map Gloves | Mar 100 Gloves | Map Goggles | Mar 100 Goggles | Map Mask | Mar 100 Mask | |:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|:------:|:---------:|:----------:|:---------:|:------:|:------:|:-------:|:---------:|:----------:|:---------:|:------------:|:----------------:|:---------------:|:-------------------:|:----------:|:--------------:|:-----------:|:---------------:|:--------:|:------------:| | No log | 1.0 | 107 | 4.7397 | 0.1272 | 0.2199 | 0.1199 | 0.0238 | 0.084...

NoteSep 27huggingface.co/nonl/dfine-cppe5

Source context: 28 downloads · 1 likes · Pipeline object-detection · Library transformers · Repo nonl/dfine-cppe5