rtdetr-r50-cppe5-finetune

This model is a fine-tuned version of PekingU/rtdetr_r50vd_coco_o365 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 10.2928
• Map: 0.5304
• Map 50: 0.7969
• Map 75: 0.5535
• Map Small: 0.5383
• Map Medium: 0.49
• Map Large: 0.6862
• Mar 1: 0.3948
• Mar 10: 0.6923
• Mar 100: 0.721
• Mar Small: 0.6578
• Mar Medium: 0.633
• Mar Large: 0.8541
• Map Coverall: 0.5634
• Mar 100 Coverall: 0.8231
• Map Face Shield: 0.5349
• Mar 100 Face Shield: 0.7765
• Map Gloves: 0.5269
• Mar 100 Gloves: 0.6169
• Map Goggles: 0.4186
• Mar 100 Goggles: 0.6414
• Map Mask: 0.6082
• Mar 100 Mask: 0.7471

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 with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 300
• num_epochs: 10

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	38.7162	0.0281	0.0511	0.0258	0.0002	0.0185	0.0643	0.0476	0.113	0.1361	0.0087	0.0487	0.2792	0.0789	0.4432	0.0001	0.0228	0.0004	0.0237	0.0152	0.0692	0.0459	0.1218
No log	2.0	214	18.9832	0.0761	0.1526	0.0644	0.0353	0.0654	0.2102	0.1353	0.3126	0.3759	0.1266	0.3259	0.6436	0.1091	0.6194	0.0156	0.3266	0.0341	0.242	0.0191	0.2846	0.2023	0.4071
No log	3.0	321	13.6077	0.1752	0.336	0.1649	0.1135	0.1322	0.3377	0.2245	0.4264	0.4943	0.3222	0.4179	0.6847	0.2821	0.6622	0.0584	0.5165	0.1717	0.4232	0.054	0.3523	0.3099	0.5173
No log	4.0	428	13.1560	0.2263	0.4207	0.2104	0.123	0.1605	0.3983	0.2559	0.4724	0.5286	0.3789	0.4497	0.7202	0.4114	0.6941	0.0924	0.5797	0.1747	0.4004	0.1153	0.4169	0.3378	0.5516
36.8374	5.0	535	12.6611	0.2505	0.4767	0.2384	0.1732	0.1978	0.4396	0.2831	0.4849	0.5407	0.4167	0.4822	0.6991	0.3467	0.691	0.1534	0.5886	0.2309	0.4187	0.1644	0.4277	0.3568	0.5773
36.8374	6.0	642	12.5659	0.2576	0.4946	0.2407	0.2101	0.2032	0.4593	0.2874	0.4959	0.5556	0.4407	0.4877	0.7199	0.3378	0.6973	0.1598	0.5987	0.2355	0.4237	0.2048	0.4785	0.3499	0.58
36.8374	7.0	749	12.4556	0.2747	0.5177	0.2548	0.2354	0.2181	0.4835	0.2967	0.4947	0.5556	0.441	0.4979	0.7231	0.3715	0.6995	0.1808	0.5861	0.2433	0.4295	0.2428	0.4769	0.3353	0.5858
36.8374	8.0	856	12.2225	0.3073	0.5745	0.2947	0.2395	0.2472	0.5148	0.3079	0.5121	0.5598	0.4353	0.491	0.726	0.4049	0.7005	0.2623	0.5924	0.2578	0.4308	0.2615	0.4923	0.3501	0.5831
36.8374	9.0	963	12.1380	0.3083	0.5642	0.3073	0.2344	0.2466	0.5222	0.3045	0.5154	0.5556	0.4475	0.4803	0.7161	0.4156	0.7005	0.2392	0.5861	0.2685	0.4366	0.2581	0.4708	0.3602	0.584
14.6693	10.0	1070	12.2646	0.3089	0.5749	0.2984	0.2304	0.257	0.519	0.3092	0.5064	0.5535	0.4401	0.4906	0.7031	0.4125	0.7009	0.2577	0.5861	0.2494	0.4295	0.2601	0.4585	0.3649	0.5924

Framework versions

• Transformers 4.54.0.dev0
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.2