msoft_detr_finetuned_cppe5_5

jaxnwagner/msoft_detr_finetuned_cppe5_5

msoft_detr_finetuned_cppe5_5 is a machine learning model from jaxnwagner. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

4% of all-time downloads

All-time downloads

257

Public

Parameters

43.5M

3.5 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors174 MB · 100%

At a glance

License: apache-2.0
Model type: conditional_detr
Access: Public
Created: Oct 21, 2024
Updated: Oct 22, 2024
SHA: 08578d7b

Base models

microsoft/conditional-detr-resnet-50

Model card

From the Hugging Face model README

msoft_detr_finetuned_cppe5_5

This model is a fine-tuned version of microsoft/conditional-detr-resnet-50 on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 1.6502
Map: 0.1506
Map 50: 0.3241
Map 75: 0.1179
Map Small: 0.0383
Map Medium: 0.1155
Map Large: 0.2123
Mar 1: 0.1856
Mar 10: 0.3707
Mar 100: 0.3953
Mar Small: 0.1845
Mar Medium: 0.3383
Mar Large: 0.5613
Map Coverall: 0.4385
Mar 100 Coverall: 0.6243
Map Face Shield: 0.0432
Mar 100 Face Shield: 0.3886
Map Gloves: 0.0719
Mar 100 Gloves: 0.3103
Map Goggles: 0.0309
Mar 100 Goggles: 0.2846
Map Mask: 0.1683
Mar 100 Mask: 0.3684

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: 20

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	50	3.3082	0.0005	0.0023	0.0	0.0005	0.0008	0.0004	0.0007	0.0115	0.0303	0.0124	0.0346	0.0328	0.0006	0.0631	0.0	0.0	0.0015	0.0531	0.0	0.0	0.0002	0.0351
No log	2.0	100	2.5406	0.0093	0.0254	0.0058	0.0083	0.0184	0.0093	0.0353	0.1003	0.1367	0.0596	0.1127	0.176	0.0281	0.2554	0.0012	0.038	0.003	0.1402	0.0012	0.0246	0.0133	0.2253
No log	3.0	150	2.2947	0.0378	0.097	0.0256	0.0102	0.0315	0.0531	0.0876	0.188	0.2363	0.1059	0.1821	0.3366	0.1268	0.4113	0.0159	0.119	0.0042	0.2004	0.0175	0.1492	0.0248	0.3013
No log	4.0	200	2.1994	0.0417	0.0962	0.0321	0.0092	0.032	0.0541	0.0946	0.2085	0.2556	0.0886	0.194	0.3656	0.1454	0.4829	0.0148	0.1165	0.0059	0.1897	0.0075	0.1585	0.0349	0.3302
No log	5.0	250	2.1116	0.0609	0.1345	0.0457	0.0149	0.0454	0.0779	0.1193	0.2433	0.2826	0.1029	0.2107	0.4202	0.2176	0.5306	0.0255	0.1861	0.011	0.2179	0.009	0.1738	0.0417	0.3044
No log	6.0	300	2.0783	0.0523	0.1158	0.043	0.0143	0.0381	0.0587	0.1021	0.2353	0.2789	0.1027	0.2118	0.3719	0.1791	0.5491	0.0137	0.1785	0.0126	0.1902	0.0116	0.1846	0.0448	0.292
No log	7.0	350	2.0252	0.0686	0.1619	0.0602	0.0155	0.051	0.0825	0.1253	0.2744	0.31	0.1124	0.2388	0.4518	0.2246	0.5617	0.0207	0.2532	0.0141	0.2326	0.0121	0.1708	0.0713	0.332
No log	8.0	400	1.9021	0.0952	0.2082	0.076	0.0155	0.0727	0.1111	0.1262	0.2869	0.3243	0.1054	0.2691	0.4416	0.3488	0.6284	0.0199	0.2114	0.0274	0.25	0.0109	0.2015	0.0689	0.3302
No log	9.0	450	1.8629	0.1124	0.2367	0.1016	0.0268	0.0809	0.1382	0.1613	0.3163	0.3522	0.1259	0.2844	0.5206	0.372	0.6162	0.0335	0.3	0.0289	0.267	0.0193	0.2262	0.1082	0.3516
3.5058	10.0	500	1.7706	0.1132	0.2447	0.0912	0.0249	0.0863	0.1384	0.1611	0.3268	0.3662	0.1789	0.3043	0.5176	0.3747	0.5905	0.0239	0.3051	0.0317	0.3009	0.0284	0.2415	0.1074	0.3929
3.5058	11.0	550	1.7552	0.1238	0.2833	0.0924	0.032	0.1008	0.1517	0.1603	0.3388	0.3762	0.1731	0.3242	0.5294	0.3793	0.5757	0.0325	0.362	0.0424	0.308	0.0275	0.2554	0.1374	0.38
3.5058	12.0	600	1.7298	0.1275	0.2938	0.0959	0.0361	0.0978	0.1648	0.1692	0.3493	0.382	0.1624	0.3309	0.5347	0.3959	0.6158	0.0438	0.381	0.0477	0.3022	0.0255	0.2662	0.1247	0.3449
3.5058	13.0	650	1.7136	0.136	0.2982	0.0999	0.0341	0.1025	0.1757	0.1758	0.3593	0.3918	0.178	0.3362	0.55	0.4242	0.6225	0.0469	0.381	0.0507	0.3022	0.0258	0.2862	0.1326	0.3671
3.5058	14.0	700	1.6856	0.1451	0.319	0.1159	0.0361	0.1075	0.1986	0.1834	0.3631	0.395	0.1736	0.3379	0.5598	0.4343	0.641	0.0486	0.381	0.0599	0.3076	0.0298	0.2754	0.1527	0.3698
3.5058	15.0	750	1.6613	0.148	0.3162	0.1197	0.0394	0.1165	0.1989	0.1836	0.3721	0.399	0.1881	0.3358	0.5663	0.4398	0.6365	0.0451	0.3962	0.0639	0.3058	0.0348	0.2877	0.1563	0.3689
3.5058	16.0	800	1.6491	0.1487	0.3267	0.1178	0.0406	0.118	0.2019	0.19	0.3722	0.3981	0.1847	0.3418	0.5637	0.4385	0.6293	0.0404	0.381	0.068	0.3098	0.0319	0.3	0.1646	0.3702
3.5058	17.0	850	1.6468	0.1489	0.3263	0.1187	0.0374	0.1188	0.2089	0.1894	0.3708	0.3978	0.1931	0.3406	0.5609	0.4355	0.6288	0.0426	0.3848	0.0693	0.3129	0.0314	0.2892	0.1658	0.3733
3.5058	18.0	900	1.6533	0.1487	0.3212	0.1177	0.0356	0.1137	0.2078	0.1856	0.3703	0.3964	0.1931	0.3356	0.5642	0.4369	0.6266	0.0412	0.3899	0.0721	0.3098	0.0292	0.2846	0.164	0.3711
3.5058	19.0	950	1.6509	0.1503	0.3241	0.1176	0.0383	0.1153	0.2121	0.1843	0.3705	0.3953	0.1848	0.3379	0.5623	0.4378	0.6234	0.0433	0.3899	0.0713	0.3103	0.0309	0.2846	0.1681	0.3684
1.4402	20.0	1000	1.6502	0.1506	0.3241	0.1179	0.0383	0.1155	0.2123	0.1856	0.3707	0.3953	0.1845	0.3383	0.5613	0.4385	0.6243	0.0432	0.3886	0.0719	0.3103	0.0309	0.2846	0.1683	0.3684

Framework versions

Transformers 4.44.0
Pytorch 2.5.0+cu124
Datasets 2.21.0
Tokenizers 0.19.1

Type: conditional_detr
License: apache-2.0
Created: Oct 21, 2024
Updated: Oct 22, 2024