yolo_finetuned_kangaroo

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

• Loss: 0.6987
• Map: 0.6588
• Map 50: 0.9376
• Map 75: 0.7247
• Map Small: -1.0
• Map Medium: 0.4086
• Map Large: 0.6921
• Mar 1: 0.65
• Mar 10: 0.8071
• Mar 100: 0.8595
• Mar Small: -1.0
• Mar Medium: 0.78
• Mar Large: 0.8703
• Map Raccoon: 0.6588
• Mar 100 Raccoon: 0.8595

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: 4
• eval_batch_size: 8
• seed: 42
• 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
• num_epochs: 30

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 Raccoon	Mar 100 Raccoon
No log	1.0	40	0.9086	0.2342	0.3563	0.2666	-1.0	0.1389	0.2586	0.4976	0.6929	0.7762	-1.0	0.5	0.8135	0.2342	0.7762
No log	2.0	80	1.0726	0.1095	0.1578	0.1314	-1.0	0.0407	0.1218	0.4167	0.6119	0.7262	-1.0	0.46	0.7622	0.1095	0.7262
No log	3.0	120	0.9048	0.1347	0.2242	0.1274	-1.0	0.0706	0.1483	0.3952	0.6071	0.7738	-1.0	0.62	0.7946	0.1347	0.7738
No log	4.0	160	0.9243	0.2736	0.4619	0.2685	-1.0	0.0772	0.3026	0.5095	0.7214	0.7571	-1.0	0.44	0.8	0.2736	0.7571
No log	5.0	200	1.0339	0.2776	0.4383	0.3325	-1.0	0.1196	0.3069	0.5405	0.6929	0.719	-1.0	0.26	0.7811	0.2776	0.719
No log	6.0	240	0.8273	0.2978	0.4797	0.2569	-1.0	0.1013	0.3282	0.5548	0.7381	0.7929	-1.0	0.62	0.8162	0.2978	0.7929
No log	7.0	280	0.8849	0.3984	0.6006	0.4514	-1.0	0.1681	0.4371	0.5381	0.7381	0.7548	-1.0	0.48	0.7919	0.3984	0.7548
No log	8.0	320	0.8198	0.4433	0.6648	0.525	-1.0	0.2698	0.4743	0.5619	0.769	0.7833	-1.0	0.5	0.8216	0.4433	0.7833
No log	9.0	360	0.7992	0.397	0.6402	0.3959	-1.0	0.3175	0.4254	0.5762	0.7452	0.7857	-1.0	0.56	0.8162	0.397	0.7857
No log	10.0	400	0.9780	0.459	0.6974	0.54	-1.0	0.1869	0.5011	0.581	0.7238	0.7333	-1.0	0.26	0.7973	0.459	0.7333
No log	11.0	440	0.8702	0.4456	0.6876	0.4884	-1.0	0.2613	0.4757	0.5667	0.7548	0.7738	-1.0	0.48	0.8135	0.4456	0.7738
No log	12.0	480	0.7883	0.5063	0.7538	0.581	-1.0	0.2773	0.5459	0.5857	0.7738	0.8095	-1.0	0.58	0.8405	0.5063	0.8095
0.7997	13.0	520	0.8627	0.526	0.7835	0.6069	-1.0	0.2573	0.5704	0.5881	0.7738	0.8024	-1.0	0.58	0.8324	0.526	0.8024
0.7997	14.0	560	0.8319	0.51	0.8101	0.5763	-1.0	0.3612	0.5364	0.5667	0.7619	0.831	-1.0	0.7	0.8486	0.51	0.831
0.7997	15.0	600	0.7695	0.5425	0.7847	0.585	-1.0	0.2498	0.5923	0.6048	0.7881	0.85	-1.0	0.66	0.8757	0.5425	0.85
0.7997	16.0	640	0.7089	0.5419	0.7761	0.5974	-1.0	0.28	0.5894	0.6071	0.7929	0.8381	-1.0	0.66	0.8622	0.5419	0.8381
0.7997	17.0	680	0.6876	0.5645	0.8054	0.6102	-1.0	0.3604	0.5949	0.6238	0.819	0.8524	-1.0	0.72	0.8703	0.5645	0.8524
0.7997	18.0	720	0.6689	0.6185	0.8905	0.695	-1.0	0.4105	0.6526	0.6405	0.8238	0.8667	-1.0	0.72	0.8865	0.6185	0.8667
0.7997	19.0	760	0.6938	0.6284	0.8981	0.6971	-1.0	0.3667	0.6634	0.6476	0.8143	0.8548	-1.0	0.74	0.8703	0.6284	0.8548
0.7997	20.0	800	0.7279	0.6261	0.8972	0.7276	-1.0	0.3391	0.664	0.6262	0.8119	0.8571	-1.0	0.74	0.873	0.6261	0.8571
0.7997	21.0	840	0.6939	0.6418	0.9111	0.7283	-1.0	0.3903	0.6745	0.6595	0.8143	0.8571	-1.0	0.74	0.873	0.6418	0.8571
0.7997	22.0	880	0.7210	0.6437	0.9314	0.7042	-1.0	0.3811	0.6764	0.631	0.8214	0.8548	-1.0	0.78	0.8649	0.6437	0.8548
0.7997	23.0	920	0.6969	0.6632	0.9398	0.7255	-1.0	0.4227	0.6943	0.6405	0.8167	0.8595	-1.0	0.78	0.8703	0.6632	0.8595
0.7997	24.0	960	0.6793	0.6576	0.9368	0.734	-1.0	0.424	0.6878	0.6476	0.8143	0.8667	-1.0	0.76	0.8811	0.6576	0.8667
0.5114	25.0	1000	0.7123	0.6612	0.926	0.7346	-1.0	0.3847	0.6999	0.6762	0.8119	0.8667	-1.0	0.74	0.8838	0.6612	0.8667
0.5114	26.0	1040	0.6958	0.6571	0.9465	0.7146	-1.0	0.3983	0.6888	0.6548	0.8119	0.8643	-1.0	0.78	0.8757	0.6571	0.8643
0.5114	27.0	1080	0.6979	0.6584	0.9456	0.7209	-1.0	0.4134	0.6898	0.6548	0.8119	0.8619	-1.0	0.78	0.873	0.6584	0.8619
0.5114	28.0	1120	0.7025	0.6563	0.9374	0.721	-1.0	0.4025	0.6891	0.6476	0.8048	0.8595	-1.0	0.78	0.8703	0.6563	0.8595
0.5114	29.0	1160	0.6985	0.659	0.9387	0.7243	-1.0	0.4091	0.6919	0.65	0.8071	0.8595	-1.0	0.78	0.8703	0.659	0.8595
0.5114	30.0	1200	0.6987	0.6588	0.9376	0.7247	-1.0	0.4086	0.6921	0.65	0.8071	0.8595	-1.0	0.78	0.8703	0.6588	0.8595

Framework versions

• Transformers 4.57.6
• Pytorch 2.9.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2