grounding-dino-tiny-aquarium-fine-tune

This model is a fine-tuned version of IDEA-Research/grounding-dino-tiny on an aquarium dataset. It achieves the following results on the evaluation set:

• Loss: 18.3797
• Map: 0.1008
• Map 50: 0.1716
• Map 75: 0.0971
• Map Small: -1.0
• Map Medium: 0.1054
• Map Large: 0.1024
• Mar 1: 0.1063
• Mar 10: 0.1758
• Mar 100: 0.2843
• Mar Small: -1.0
• Mar Medium: 0.1848
• Mar Large: 0.2881
• Map Fish: 0.0827
• Mar 100 Fish: 0.411
• Map Jellyfish: 0.1291
• Mar 100 Jellyfish: 0.4026
• Map Penguins: 0.0963
• Mar 100 Penguins: 0.426
• Map Sharks: 0.0336
• Mar 100 Sharks: 0.0561
• Map Puffins: 0.04
• Mar 100 Puffins: 0.1
• Map Stingrays: 0.2407
• Mar 100 Stingrays: 0.4042
• Map Starfish: 0.0829
• Mar 100 Starfish: 0.1906

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: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• 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 Fish	Mar 100 Fish	Map Jellyfish	Mar 100 Jellyfish	Map Penguins	Mar 100 Penguins	Map Sharks	Mar 100 Sharks	Map Puffins	Mar 100 Puffins	Map Stingrays	Mar 100 Stingrays	Map Starfish	Mar 100 Starfish
47786.9107	1.0	112	47816.1406	0.0006	0.0019	0.0003	-1.0	0.0012	0.0008	0.0005	0.0071	0.0116	-1.0	0.0455	0.0107	0.0028	0.0491	0.0017	0.0323	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
36235.5402	2.0	224	26297.4707	0.0036	0.0117	0.0021	-1.0	0.0002	0.0039	0.0014	0.0092	0.0326	-1.0	0.0119	0.0341	0.0182	0.1514	0.0069	0.0768	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
13131.6674	3.0	336	4944.3315	0.0022	0.0039	0.0016	-1.0	0.0005	0.0022	0.0017	0.005	0.0085	-1.0	0.0107	0.0087	0.0139	0.0546	0.0013	0.0052	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
1436.2895	4.0	448	160.7001	0.0065	0.0175	0.004	-1.0	0.0074	0.0068	0.0044	0.0129	0.0322	-1.0	0.0119	0.0336	0.0202	0.1621	0.0254	0.0632	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
48.6727	5.0	560	19.5311	0.0183	0.0474	0.0105	-1.0	0.0086	0.0197	0.0081	0.0338	0.0768	-1.0	0.0399	0.0798	0.0332	0.2379	0.0663	0.211	0.0288	0.0885	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
19.4237	6.0	672	18.9543	0.0649	0.12	0.0633	-1.0	0.1113	0.0658	0.0817	0.1352	0.2257	-1.0	0.1698	0.2284	0.0611	0.3724	0.0829	0.3542	0.0444	0.3154	0.0	0.0	0.0113	0.0574	0.202	0.3583	0.0529	0.1219
18.6905	7.0	784	18.5298	0.0855	0.1458	0.0805	-1.0	0.0869	0.087	0.109	0.1788	0.2767	-1.0	0.165	0.2801	0.0615	0.3822	0.1007	0.3826	0.0772	0.3538	0.0107	0.0258	0.02	0.0852	0.2416	0.4792	0.0871	0.2281
18.3457	8.0	896	18.6190	0.0808	0.1485	0.0724	-1.0	0.1232	0.0818	0.1071	0.1821	0.2959	-1.0	0.201	0.2993	0.0644	0.4116	0.1096	0.4142	0.0761	0.4067	0.0208	0.0833	0.0386	0.1574	0.2184	0.4542	0.0379	0.1437
18.2089	9.0	1008	18.3819	0.0997	0.1662	0.0979	-1.0	0.0888	0.1015	0.098	0.1657	0.2626	-1.0	0.1938	0.2653	0.0726	0.3886	0.1283	0.3903	0.0993	0.3904	0.0207	0.05	0.0236	0.0593	0.2596	0.425	0.094	0.1344
17.7603	10.0	1120	18.3797	0.1008	0.1716	0.0971	-1.0	0.1054	0.1024	0.1063	0.1758	0.2843	-1.0	0.1848	0.2881	0.0827	0.411	0.1291	0.4026	0.0963	0.426	0.0336	0.0561	0.04	0.1	0.2407	0.4042	0.0829	0.1906

Framework versions

• Transformers 4.45.0.dev0
• Pytorch 2.1.2
• Datasets 2.20.0
• Tokenizers 0.19.1