johanbkg2513/detr_resnet50_finetuned_trashify_object_detector_v1

来源：🤗🏴‍☠️ 救援者：isiliconobject-detectionapache-2.042M 参数167 MBsafetensors✓ 2 个校验和167 天前更新

detr_resnet50_finetuned_trashify_box_detector_v1

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: 1.2899
Map: 0.1934
Map 50: 0.306
Map 75: 0.2065
Map Small: 0.0
Map Medium: 0.0689
Map Large: 0.2001
Mar 1: 0.2433
Mar 10: 0.3701
Mar 100: 0.3903
Mar Small: 0.0
Mar Medium: 0.1534
Mar Large: 0.4068
Map Bin: 0.514
Mar 100 Bin: 0.7837
Map Hand: 0.2815
Mar 100 Hand: 0.6304
Map Not Bin: 0.0
Mar 100 Not Bin: 0.0
Map Not Trash: 0.0374
Mar 100 Not Trash: 0.3417
Map Trash: 0.3272
Mar 100 Trash: 0.5858
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: linear
lr_scheduler_warmup_steps: 0.05
num_epochs: 10
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Map	Map 50	Map 75	Map Medium	Map Large	Mar 1	Mar 10	Mar 100	Mar Medium	Mar Large	Map Bin	Mar 100 Bin	Map Hand	Mar 100 Hand	Map Not Trash	Mar 100 Not Trash	Map Trash	Mar 100 Trash
2.6558	1.0	50	2.1875	0.0126	0.0276	0.0094	0.0	0.0128	0.026	0.076	0.1259	0.0	0.1271	0.067	0.6929	0.0085	0.0559	0.0001	0.0056	0.0001	0.0009
2.0859	2.0	100	1.9046	0.0621	0.1167	0.0607	0.0382	0.0623	0.1299	0.2344	0.2706	0.0784	0.2767	0.1862	0.7163	0.0955	0.4402	0.0033	0.1306	0.0876	0.3363
1.8376	3.0	150	1.6975	0.0883	0.156	0.0896	0.0332	0.0895	0.1598	0.2986	0.3253	0.0989	0.3353	0.2473	0.6908	0.1524	0.5775	0.0045	0.1861	0.1258	0.4973
1.6784	4.0	200	1.6278	0.1093	0.1897	0.1155	0.0245	0.1145	0.1825	0.2926	0.3078	0.0886	0.3179	0.2566	0.6943	0.2266	0.5284	0.0078	0.1889	0.1649	0.4354
1.7628	5.0	250	1.6699	0.1096	0.1985	0.1102	0.0318	0.1124	0.1819	0.2892	0.3044	0.1148	0.3136	0.2831	0.6496	0.2077	0.5373	0.0136	0.1972	0.1533	0.4425
1.5951	6.0	300	1.4920	0.1538	0.2613	0.1664	0.0527	0.1577	0.2165	0.337	0.3695	0.1409	0.3837	0.3898	0.7702	0.2573	0.5873	0.0221	0.2694	0.2535	0.5903
1.4962	7.0	350	1.4569	0.159	0.2654	0.1724	0.0691	0.1624	0.2203	0.351	0.3708	0.142	0.3847	0.3957	0.7738	0.2413	0.602	0.0346	0.2556	0.2823	0.5938
1.4151	8.0	400	1.3825	0.1798	0.2917	0.2025	0.0744	0.1844	0.2377	0.3582	0.3709	0.1648	0.3814	0.4823	0.7681	0.2737	0.6118	0.0332	0.2625	0.2897	0.5832
1.3674	9.0	450	1.3161	0.1828	0.2983	0.1977	0.0611	0.1888	0.2397	0.367	0.3805	0.1398	0.3971	0.4858	0.778	0.2744	0.6284	0.0335	0.3083	0.3029	0.5681
1.2989	10.0	500	1.2899	0.1934	0.306	0.2065	0.0689	0.2001	0.2433	0.3701	0.3903	0.1534	0.4068	0.514	0.7837	0.2815	0.6304	0.0374	0.3417	0.3272	0.5858

Framework versions

Transformers 5.0.0
Pytorch 2.11.0+cu128
Datasets 4.8.5
Tokenizers 0.22.2