dinov2-base-imagenet1k-1-layer-finetuned-galaxy10-decals-head-finetuned-100-galaxy_mnist

This model is a fine-tuned version of matthieulel/dinov2-base-imagenet1k-1-layer-finetuned-galaxy10-decals on the matthieulel/galaxy_mnist dataset. It achieves the following results on the evaluation set:

• Loss: 0.2235
• Accuracy: 0.909
• Precision: 0.9108
• Recall: 0.909
• F1: 0.9089

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• 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
1.1235	0.99	62	0.9579	0.622	0.6410	0.622	0.6249
0.5488	2.0	125	0.3964	0.852	0.8534	0.852	0.8523
0.384	2.99	187	0.2877	0.882	0.8820	0.882	0.8819
0.3493	4.0	250	0.2621	0.8885	0.8885	0.8885	0.8885
0.3184	4.99	312	0.2570	0.8945	0.8957	0.8945	0.8946
0.3027	6.0	375	0.2485	0.9	0.9013	0.9	0.9000
0.3245	6.99	437	0.2397	0.9035	0.9038	0.9035	0.9034
0.2666	8.0	500	0.2413	0.9035	0.9051	0.9035	0.9034
0.2705	8.99	562	0.2336	0.905	0.9057	0.905	0.9050
0.2524	10.0	625	0.2320	0.9065	0.9072	0.9065	0.9065
0.271	10.99	687	0.2293	0.908	0.9085	0.908	0.9079
0.2624	12.0	750	0.2266	0.9045	0.9047	0.9045	0.9045
0.2681	12.99	812	0.2286	0.908	0.9088	0.908	0.9079
0.262	14.0	875	0.2239	0.9085	0.9089	0.9085	0.9085
0.3015	14.99	937	0.2265	0.9075	0.9085	0.9075	0.9074
0.2635	16.0	1000	0.2256	0.9065	0.9082	0.9065	0.9064
0.2708	16.99	1062	0.2239	0.9075	0.9089	0.9075	0.9075
0.2616	18.0	1125	0.2231	0.9075	0.9084	0.9075	0.9074
0.2658	18.99	1187	0.2229	0.9075	0.9092	0.9075	0.9075
0.2578	20.0	1250	0.2216	0.908	0.9093	0.908	0.9080
0.2823	20.99	1312	0.2206	0.9075	0.9084	0.9075	0.9075
0.2714	22.0	1375	0.2223	0.9085	0.9100	0.9085	0.9084
0.2631	22.99	1437	0.2235	0.909	0.9108	0.909	0.9089
0.2695	24.0	1500	0.2214	0.9085	0.9101	0.9085	0.9084
0.2698	24.99	1562	0.2208	0.908	0.9095	0.908	0.9079
0.2285	26.0	1625	0.2188	0.9075	0.9083	0.9075	0.9075
0.2716	26.99	1687	0.2202	0.9075	0.9089	0.9075	0.9075
0.2628	28.0	1750	0.2205	0.9075	0.9090	0.9075	0.9074
0.2619	28.99	1812	0.2207	0.907	0.9085	0.907	0.9069
0.2568	29.76	1860	0.2204	0.907	0.9085	0.907	0.9069

Framework versions

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