accident-detection-model

This model is a fine-tuned version of PekingU/rtdetr_v2_r50vd on the None dataset. It achieves the following results on the evaluation set:

• Loss: 9.4502
• Map: 0.5515
• Map 50: 0.7683
• Map 75: 0.6079
• Map Small: 0.0994
• Map Medium: 0.4989
• Map Large: 0.6537
• Mar 1: 0.4326
• Mar 10: 0.6527
• Mar 100: 0.7307
• Mar Small: 0.2089
• Mar Medium: 0.6685
• Mar Large: 0.8355
• Map Accident: 0.6888
• Mar 100 Accident: 0.8481
• Map Vehicle: 0.4142
• Mar 100 Vehicle: 0.6132

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: 0.0001
• train_batch_size: 4
• eval_batch_size: 4
• seed: 3407
• 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: 300
• num_epochs: 15
• mixed_precision_training: Native AMP

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 Accident	Mar 100 Accident	Map Vehicle	Mar 100 Vehicle
27.5163	1.0	166	14.6379	0.2815	0.4168	0.3063	0.0865	0.1986	0.3439	0.3178	0.6023	0.7048	0.1912	0.5613	0.8163	0.2561	0.8009	0.3069	0.6086
20.6308	2.0	332	13.4814	0.2503	0.4052	0.2656	0.0771	0.194	0.2896	0.2346	0.472	0.6041	0.194	0.4528	0.6835	0.2435	0.6696	0.257	0.5386
19.2242	3.0	498	11.2376	0.4394	0.646	0.4938	0.091	0.4047	0.5039	0.3577	0.6223	0.701	0.2099	0.6207	0.8112	0.5285	0.8009	0.3503	0.601
18.0153	4.0	664	11.0048	0.4124	0.5958	0.4506	0.071	0.3778	0.4667	0.3692	0.6101	0.7151	0.2089	0.6353	0.8216	0.5426	0.8276	0.2823	0.6026
15.3270	5.0	830	10.3142	0.4482	0.6293	0.4854	0.0767	0.3959	0.5393	0.3837	0.614	0.7021	0.2014	0.6305	0.8033	0.5768	0.8131	0.3196	0.591
15.1404	6.0	996	10.3167	0.48	0.6925	0.5128	0.0926	0.3978	0.5631	0.3694	0.6394	0.712	0.2154	0.6134	0.816	0.5747	0.8093	0.3853	0.6146
15.2354	7.0	1162	9.8965	0.4989	0.7199	0.5461	0.0957	0.4056	0.6025	0.4015	0.6438	0.7244	0.2165	0.6446	0.8224	0.591	0.8276	0.4068	0.6212
13.1620	8.0	1328	9.8355	0.5224	0.7497	0.5783	0.1009	0.4833	0.6154	0.4113	0.65	0.7346	0.2175	0.6701	0.8348	0.6375	0.8463	0.4074	0.623
12.6002	9.0	1494	9.7019	0.5418	0.7604	0.5956	0.1007	0.479	0.6413	0.4185	0.6517	0.7325	0.2214	0.6451	0.8314	0.6624	0.835	0.4212	0.63
11.6298	10.0	1660	9.6623	0.5376	0.7582	0.5915	0.1097	0.4634	0.634	0.4183	0.6635	0.7388	0.2249	0.6913	0.837	0.6572	0.8435	0.418	0.6342
11.0523	11.0	1826	9.4776	0.5504	0.7645	0.6063	0.0973	0.4954	0.6617	0.4315	0.6565	0.7312	0.2074	0.649	0.844	0.6908	0.8491	0.4101	0.6132
10.7115	12.0	1992	9.5361	0.5474	0.763	0.6048	0.0955	0.5085	0.6465	0.4307	0.6486	0.7309	0.2078	0.6831	0.8315	0.6848	0.8467	0.4099	0.6151
10.1611	13.0	2158	9.4691	0.5508	0.7672	0.6031	0.1017	0.495	0.6505	0.4322	0.6588	0.7337	0.2105	0.6807	0.8378	0.6884	0.8505	0.4132	0.617
10.6499	14.0	2324	9.4728	0.551	0.7714	0.6092	0.1007	0.5017	0.6544	0.4298	0.6553	0.7338	0.2105	0.6815	0.8388	0.6851	0.8491	0.4169	0.6185
10.0297	15.0	2490	9.4502	0.5515	0.7683	0.6079	0.0994	0.4989	0.6537	0.4326	0.6527	0.7307	0.2089	0.6685	0.8355	0.6888	0.8481	0.4142	0.6132

Framework versions

• Transformers 5.15.1
• Pytorch 2.11.0+cu128
• Datasets 5.0.1
• Tokenizers 0.22.2