dinov2-base-imagenet1k-1-layer-finetuned-galaxy10-decals

This model is a fine-tuned version of facebook/dinov2-base-imagenet1k-1-layer on the matthieulel/galaxy10_decals dataset. It achieves the following results on the evaluation set:

• Loss: 0.5267
• Accuracy: 0.8670
• Precision: 0.8645
• Recall: 0.8670
• F1: 0.8650

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: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.8157	0.99	62	0.6740	0.7813	0.8046	0.7813	0.7853
0.8091	2.0	125	0.5948	0.8021	0.8016	0.8021	0.7950
0.6983	2.99	187	0.6016	0.7965	0.8077	0.7965	0.7909
0.6701	4.0	250	0.5676	0.7982	0.8016	0.7982	0.7954
0.5998	4.99	312	0.5116	0.8286	0.8401	0.8286	0.8302
0.5521	6.0	375	0.5155	0.8354	0.8375	0.8354	0.8325
0.5441	6.99	437	0.5574	0.8033	0.8104	0.8033	0.7980
0.5142	8.0	500	0.4818	0.8410	0.8418	0.8410	0.8376
0.5136	8.99	562	0.4914	0.8337	0.8353	0.8337	0.8317
0.4533	10.0	625	0.4740	0.8320	0.8335	0.8320	0.8295
0.4904	10.99	687	0.5075	0.8399	0.8409	0.8399	0.8375
0.4361	12.0	750	0.4552	0.8563	0.8554	0.8563	0.8540
0.414	12.99	812	0.5025	0.8365	0.8455	0.8365	0.8374
0.4114	14.0	875	0.4822	0.8467	0.8437	0.8467	0.8420
0.3878	14.99	937	0.4615	0.8574	0.8552	0.8574	0.8549
0.3756	16.0	1000	0.5017	0.8444	0.8523	0.8444	0.8449
0.3056	16.99	1062	0.4910	0.8517	0.8495	0.8517	0.8501
0.3255	18.0	1125	0.5206	0.8523	0.8505	0.8523	0.8491
0.3224	18.99	1187	0.5066	0.8450	0.8470	0.8450	0.8438
0.2763	20.0	1250	0.5043	0.8574	0.8519	0.8574	0.8534
0.2926	20.99	1312	0.5345	0.8546	0.8542	0.8546	0.8512
0.2824	22.0	1375	0.5320	0.8529	0.8523	0.8529	0.8517
0.2613	22.99	1437	0.5254	0.8563	0.8543	0.8563	0.8542
0.2292	24.0	1500	0.5553	0.8546	0.8529	0.8546	0.8528
0.2313	24.99	1562	0.5603	0.8602	0.8612	0.8602	0.8593
0.2143	26.0	1625	0.5267	0.8670	0.8645	0.8670	0.8650
0.2075	26.99	1687	0.5737	0.8574	0.8589	0.8574	0.8573
0.2121	28.0	1750	0.5748	0.8619	0.8601	0.8619	0.8604
0.1944	28.99	1812	0.5666	0.8647	0.8618	0.8647	0.8624
0.1866	29.76	1860	0.5676	0.8608	0.8583	0.8608	0.8589

Framework versions

• Transformers 4.37.2
• Pytorch 2.3.0
• Datasets 2.19.1
• Tokenizers 0.15.1