detr_resnet50_finetuned_trashify_box_detector_v2

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

Loss: 1.0192
Map: 0.2987
Map 50: 0.4486
Map 75: 0.333
Map Small: 0.05
Map Medium: 0.2194
Map Large: 0.3138
Mar 1: 0.3228
Mar 10: 0.4392
Mar 100: 0.4556
Mar Small: 0.05
Mar Medium: 0.2943
Mar Large: 0.4829
Map Bin: 0.669
Mar 100 Bin: 0.8206
Map Hand: 0.4642
Mar 100 Hand: 0.7363
Map Not Bin: 0.0
Mar 100 Not Bin: 0.0
Map Not Trash: 0.152
Mar 100 Not Trash: 0.5292
Map Trash: 0.5071
Mar 100 Trash: 0.6478
Map Trash Arm: 0.0
Mar 100 Trash Arm: 0.0

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: 42
gradient_accumulation_steps: 4
total_train_batch_size: 16
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: 20
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 Bin	Mar 100 Bin	Map Hand	Mar 100 Hand	Map Not Bin	Mar 100 Not Bin	Map Not Hand	Mar 100 Not Hand	Map Not Trash	Mar 100 Not Trash	Map Trash	Mar 100 Trash	Map Trash Arm	Mar 100 Trash Arm
3.1541	1.0	50	2.0485	0.0148	0.0303	0.0131	0.0	0.0132	0.0173	0.0677	0.1892	0.2625	0.0	0.0545	0.2748	0.0273	0.6922	0.0202	0.3775	0.0026	0.0571	-1.0	-1.0	0.0078	0.0458	0.0309	0.4027	0.0	0.0
1.9247	2.0	100	1.6989	0.0733	0.1342	0.0718	0.0	0.0158	0.0767	0.2014	0.2905	0.3233	0.0	0.0432	0.3338	0.1322	0.7766	0.1822	0.5873	0.0	0.0	0.0137	0.0222	0.1116	0.554	0.0	0.0
1.7924	3.0	150	1.4695	0.1009	0.1668	0.1104	0.0	0.0497	0.1046	0.2316	0.3458	0.3807	0.0	0.1034	0.393	0.1869	0.8071	0.1887	0.6804	0.0	0.0	0.0178	0.1903	0.212	0.6062	0.0	0.0
1.5961	4.0	200	1.3696	0.1344	0.2258	0.1371	0.15	0.0655	0.1388	0.2312	0.3668	0.3884	0.15	0.1375	0.404	0.2912	0.7688	0.2401	0.6735	0.0	0.0	0.0171	0.3111	0.258	0.577	0.0	0.0
1.4867	5.0	250	1.2413	0.1961	0.313	0.2142	0.25	0.0655	0.2051	0.2588	0.3958	0.4122	0.25	0.1455	0.4349	0.4457	0.7702	0.3067	0.6912	0.0	0.0	0.0367	0.3889	0.3878	0.623	0.0	0.0
1.4197	6.0	300	1.2276	0.228	0.3749	0.2324	0.05	0.0556	0.2401	0.2763	0.3876	0.3976	0.15	0.1091	0.4208	0.487	0.7355	0.379	0.701	0.0	0.0	0.0642	0.3625	0.4377	0.5867	0.0	0.0
1.3595	7.0	350	1.1540	0.2221	0.3777	0.2288	0.175	0.0725	0.235	0.2627	0.3975	0.4145	0.25	0.1398	0.4351	0.5201	0.7957	0.3424	0.7333	0.0	0.0	0.0631	0.3958	0.4071	0.5619	0.0	0.0
1.3064	8.0	400	1.1456	0.2439	0.3921	0.2812	0.3	0.0706	0.2583	0.2786	0.4072	0.4226	0.3	0.1443	0.4462	0.5487	0.7582	0.3595	0.7088	0.0	0.0	0.0959	0.4569	0.4594	0.6115	0.0	0.0
1.2413	9.0	450	1.1302	0.2523	0.3928	0.2893	0.05	0.0823	0.2663	0.2997	0.4143	0.4341	0.05	0.1477	0.4636	0.5832	0.7738	0.3808	0.7225	0.0	0.0	0.0854	0.4694	0.4642	0.6389	0.0	0.0
1.2340	10.0	500	1.1006	0.2522	0.4056	0.2773	0.25	0.1886	0.2657	0.2893	0.418	0.4374	0.25	0.2807	0.461	0.596	0.7929	0.347	0.7069	0.0	0.0	0.1054	0.4917	0.465	0.6327	0.0	0.0
1.2048	11.0	550	1.0956	0.2694	0.4247	0.3005	0.05	0.1616	0.2813	0.3097	0.4193	0.4384	0.05	0.2307	0.4668	0.6181	0.7801	0.4131	0.7216	0.0	0.0	0.0936	0.4833	0.4917	0.6451	0.0	0.0
1.1786	12.0	600	1.0536	0.283	0.4403	0.3026	0.25	0.2111	0.2958	0.3128	0.4255	0.4431	0.25	0.2682	0.4707	0.6465	0.8106	0.4295	0.7137	0.0	0.0	0.123	0.5	0.4989	0.6345	0.0	0.0
1.1576	13.0	650	1.0315	0.2796	0.4265	0.3137	0.0	0.1397	0.294	0.3155	0.4307	0.4482	0.0	0.2193	0.4727	0.6253	0.7936	0.4267	0.7412	0.0	0.0	0.131	0.5056	0.4946	0.6487	0.0	0.0
1.1370	14.0	700	1.0475	0.2915	0.4432	0.3226	0.0	0.2205	0.3068	0.3194	0.4327	0.4462	0.0	0.2955	0.4722	0.6414	0.7943	0.4507	0.7206	0.0	0.0	0.1409	0.5222	0.5159	0.6398	0.0	0.0
1.1443	15.0	750	1.0357	0.2989	0.4495	0.3332	0.05	0.2035	0.3142	0.3276	0.4398	0.454	0.05	0.2739	0.4812	0.6584	0.8106	0.4662	0.7324	0.0	0.0	0.1535	0.5306	0.515	0.6504	0.0	0.0
1.1314	16.0	800	1.0217	0.2987	0.4512	0.3338	0.0	0.2095	0.3151	0.325	0.4358	0.4485	0.0	0.2818	0.4756	0.6665	0.8177	0.4719	0.7304	0.0	0.0	0.1594	0.5111	0.4945	0.6319	0.0	0.0
1.1033	17.0	850	1.0223	0.3	0.451	0.334	0.05	0.2186	0.3145	0.3267	0.4399	0.4547	0.05	0.2909	0.4821	0.6612	0.8099	0.4762	0.751	0.0	0.0	0.1582	0.5264	0.5041	0.6407	0.0	0.0
1.1038	18.0	900	1.0220	0.2993	0.4486	0.3347	0.0	0.2115	0.3142	0.3249	0.4393	0.4555	0.0	0.2898	0.4834	0.6691	0.8213	0.4673	0.7402	0.0	0.0	0.1527	0.5292	0.5067	0.6425	0.0	0.0
1.0841	19.0	950	1.0209	0.298	0.4485	0.3329	0.0	0.222	0.3129	0.3222	0.4392	0.4562	0.0	0.2977	0.4832	0.6658	0.8213	0.466	0.7392	0.0	0.0	0.1502	0.5278	0.5058	0.6487	0.0	0.0
1.0858	20.0	1000	1.0192	0.2987	0.4486	0.333	0.05	0.2194	0.3138	0.3228	0.4392	0.4556	0.05	0.2943	0.4829	0.669	0.8206	0.4642	0.7363	0.0	0.0	0.152	0.5292	0.5071	0.6478	0.0	0.0

Framework versions

Transformers 5.20.1
Pytorch 2.11.0+cu130
Datasets 5.1.0
Tokenizers 0.23.3

加载中…

detr_resnet50_finetuned_trashify_box_detector_v2

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

Loss: 1.0192

Map: 0.2987

Map 50: 0.4486

Map 75: 0.333

Map Small: 0.05

Map Medium: 0.2194

Map Large: 0.3138

Mar 1: 0.3228

Mar 10: 0.4392

Mar 100: 0.4556

Mar Small: 0.05

Mar Medium: 0.2943

Mar Large: 0.4829

Map Bin: 0.669

Mar 100 Bin: 0.8206

Map Hand: 0.4642

Mar 100 Hand: 0.7363

Map Not Bin: 0.0

Mar 100 Not Bin: 0.0

Map Not Trash: 0.152

Mar 100 Not Trash: 0.5292

Map Trash: 0.5071

Mar 100 Trash: 0.6478

Map Trash Arm: 0.0

Mar 100 Trash Arm: 0.0

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 0.0001

train_batch_size: 4

eval_batch_size: 4

seed: 42

gradient_accumulation_steps: 4

total_train_batch_size: 16

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: 20

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 Bin	Mar 100 Bin	Map Hand	Mar 100 Hand	Map Not Bin	Mar 100 Not Bin	Map Not Hand	Mar 100 Not Hand	Map Not Trash	Mar 100 Not Trash	Map Trash	Mar 100 Trash	Map Trash Arm	Mar 100 Trash Arm
3.1541	1.0	50	2.0485	0.0148	0.0303	0.0131	0.0	0.0132	0.0173	0.0677	0.1892	0.2625	0.0	0.0545	0.2748	0.0273	0.6922	0.0202	0.3775	0.0026	0.0571	-1.0	-1.0	0.0078	0.0458	0.0309	0.4027	0.0	0.0
1.9247	2.0	100	1.6989	0.0733	0.1342	0.0718	0.0	0.0158	0.0767	0.2014	0.2905	0.3233	0.0	0.0432	0.3338	0.1322	0.7766	0.1822	0.5873	0.0	0.0	0.0137	0.0222	0.1116	0.554	0.0	0.0
1.7924	3.0	150	1.4695	0.1009	0.1668	0.1104	0.0	0.0497	0.1046	0.2316	0.3458	0.3807	0.0	0.1034	0.393	0.1869	0.8071	0.1887	0.6804	0.0	0.0	0.0178	0.1903	0.212	0.6062	0.0	0.0
1.5961	4.0	200	1.3696	0.1344	0.2258	0.1371	0.15	0.0655	0.1388	0.2312	0.3668	0.3884	0.15	0.1375	0.404	0.2912	0.7688	0.2401	0.6735	0.0	0.0	0.0171	0.3111	0.258	0.577	0.0	0.0
1.4867	5.0	250	1.2413	0.1961	0.313	0.2142	0.25	0.0655	0.2051	0.2588	0.3958	0.4122	0.25	0.1455	0.4349	0.4457	0.7702	0.3067	0.6912	0.0	0.0	0.0367	0.3889	0.3878	0.623	0.0	0.0
1.4197	6.0	300	1.2276	0.228	0.3749	0.2324	0.05	0.0556	0.2401	0.2763	0.3876	0.3976	0.15	0.1091	0.4208	0.487	0.7355	0.379	0.701	0.0	0.0	0.0642	0.3625	0.4377	0.5867	0.0	0.0
1.3595	7.0	350	1.1540	0.2221	0.3777	0.2288	0.175	0.0725	0.235	0.2627	0.3975	0.4145	0.25	0.1398	0.4351	0.5201	0.7957	0.3424	0.7333	0.0	0.0	0.0631	0.3958	0.4071	0.5619	0.0	0.0
1.3064	8.0	400	1.1456	0.2439	0.3921	0.2812	0.3	0.0706	0.2583	0.2786	0.4072	0.4226	0.3	0.1443	0.4462	0.5487	0.7582	0.3595	0.7088	0.0	0.0	0.0959	0.4569	0.4594	0.6115	0.0	0.0
1.2413	9.0	450	1.1302	0.2523	0.3928	0.2893	0.05	0.0823	0.2663	0.2997	0.4143	0.4341	0.05	0.1477	0.4636	0.5832	0.7738	0.3808	0.7225	0.0	0.0	0.0854	0.4694	0.4642	0.6389	0.0	0.0
1.2340	10.0	500	1.1006	0.2522	0.4056	0.2773	0.25	0.1886	0.2657	0.2893	0.418	0.4374	0.25	0.2807	0.461	0.596	0.7929	0.347	0.7069	0.0	0.0	0.1054	0.4917	0.465	0.6327	0.0	0.0
1.2048	11.0	550	1.0956	0.2694	0.4247	0.3005	0.05	0.1616	0.2813	0.3097	0.4193	0.4384	0.05	0.2307	0.4668	0.6181	0.7801	0.4131	0.7216	0.0	0.0	0.0936	0.4833	0.4917	0.6451	0.0	0.0
1.1786	12.0	600	1.0536	0.283	0.4403	0.3026	0.25	0.2111	0.2958	0.3128	0.4255	0.4431	0.25	0.2682	0.4707	0.6465	0.8106	0.4295	0.7137	0.0	0.0	0.123	0.5	0.4989	0.6345	0.0	0.0
1.1576	13.0	650	1.0315	0.2796	0.4265	0.3137	0.0	0.1397	0.294	0.3155	0.4307	0.4482	0.0	0.2193	0.4727	0.6253	0.7936	0.4267	0.7412	0.0	0.0	0.131	0.5056	0.4946	0.6487	0.0	0.0
1.1370	14.0	700	1.0475	0.2915	0.4432	0.3226	0.0	0.2205	0.3068	0.3194	0.4327	0.4462	0.0	0.2955	0.4722	0.6414	0.7943	0.4507	0.7206	0.0	0.0	0.1409	0.5222	0.5159	0.6398	0.0	0.0
1.1443	15.0	750	1.0357	0.2989	0.4495	0.3332	0.05	0.2035	0.3142	0.3276	0.4398	0.454	0.05	0.2739	0.4812	0.6584	0.8106	0.4662	0.7324	0.0	0.0	0.1535	0.5306	0.515	0.6504	0.0	0.0
1.1314	16.0	800	1.0217	0.2987	0.4512	0.3338	0.0	0.2095	0.3151	0.325	0.4358	0.4485	0.0	0.2818	0.4756	0.6665	0.8177	0.4719	0.7304	0.0	0.0	0.1594	0.5111	0.4945	0.6319	0.0	0.0
1.1033	17.0	850	1.0223	0.3	0.451	0.334	0.05	0.2186	0.3145	0.3267	0.4399	0.4547	0.05	0.2909	0.4821	0.6612	0.8099	0.4762	0.751	0.0	0.0	0.1582	0.5264	0.5041	0.6407	0.0	0.0
1.1038	18.0	900	1.0220	0.2993	0.4486	0.3347	0.0	0.2115	0.3142	0.3249	0.4393	0.4555	0.0	0.2898	0.4834	0.6691	0.8213	0.4673	0.7402	0.0	0.0	0.1527	0.5292	0.5067	0.6425	0.0	0.0
1.0841	19.0	950	1.0209	0.298	0.4485	0.3329	0.0	0.222	0.3129	0.3222	0.4392	0.4562	0.0	0.2977	0.4832	0.6658	0.8213	0.466	0.7392	0.0	0.0	0.1502	0.5278	0.5058	0.6487	0.0	0.0
1.0858	20.0	1000	1.0192	0.2987	0.4486	0.333	0.05	0.2194	0.3138	0.3228	0.4392	0.4556	0.05	0.2943	0.4829	0.669	0.8206	0.4642	0.7363	0.0	0.0	0.152	0.5292	0.5071	0.6478	0.0	0.0

Framework versions

Transformers 5.20.1

Pytorch 2.11.0+cu130

Datasets 5.1.0

Tokenizers 0.23.3

johanbkg2513/detr_resnet50_finetuned_trashify_box_detector_v2

detr_resnet50_finetuned_trashify_box_detector_v2

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

detr_resnet50_finetuned_trashify_box_detector_v2

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions