This model is a fine-tuned version of google/vit-huge-patch14-224-in21k on the matthieulel/galaxy10_decals dataset. It achieves the following results on the evaluation set:

Loss: 0.4927
Accuracy: 0.8523
Precision: 0.8538
Recall: 0.8523
F1: 0.8489

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: 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
1.7563	0.99	62	1.6011	0.5096	0.4694	0.5096	0.4415
1.0516	2.0	125	0.9115	0.7661	0.7679	0.7661	0.7525
0.8551	2.99	187	0.7590	0.7706	0.7860	0.7706	0.7653
0.6701	4.0	250	0.6253	0.8095	0.8013	0.8095	0.7985
0.6112	4.99	312	0.6058	0.8095	0.8120	0.8095	0.8083
0.6109	6.0	375	0.5428	0.8292	0.8353	0.8292	0.8196
0.5643	6.99	437	0.5230	0.8343	0.8350	0.8343	0.8332
0.5204	8.0	500	0.5010	0.8365	0.8391	0.8365	0.8344
0.4918	8.99	562	0.5000	0.8365	0.8419	0.8365	0.8348
0.4673	10.0	625	0.4949	0.8410	0.8394	0.8410	0.8371
0.4569	10.99	687	0.4803	0.8467	0.8451	0.8467	0.8446
0.4164	12.0	750	0.5012	0.8326	0.8314	0.8326	0.8295
0.424	12.99	812	0.4940	0.8410	0.8454	0.8410	0.8382
0.4045	14.0	875	0.4927	0.8523	0.8538	0.8523	0.8489
0.3651	14.99	937	0.4809	0.8416	0.8396	0.8416	0.8403
0.3512	16.0	1000	0.4955	0.8331	0.8306	0.8331	0.8307
0.2922	16.99	1062	0.5103	0.8399	0.8357	0.8399	0.8359
0.3212	18.0	1125	0.5197	0.8439	0.8408	0.8439	0.8412
0.3171	18.99	1187	0.5253	0.8348	0.8335	0.8348	0.8335
0.2896	20.0	1250	0.5303	0.8467	0.8456	0.8467	0.8438
0.271	20.99	1312	0.5571	0.8393	0.8391	0.8393	0.8366
0.2996	22.0	1375	0.5468	0.8422	0.8411	0.8422	0.8404
0.2663	22.99	1437	0.5620	0.8405	0.8393	0.8405	0.8393
0.2513	24.0	1500	0.5338	0.8467	0.8448	0.8467	0.8450
0.2453	24.99	1562	0.5562	0.8484	0.8452	0.8484	0.8446
0.2237	26.0	1625	0.5619	0.8467	0.8450	0.8467	0.8442
0.2296	26.99	1687	0.5751	0.8484	0.8496	0.8484	0.8464
0.2479	28.0	1750	0.5782	0.8461	0.8441	0.8461	0.8431
0.2207	28.99	1812	0.5746	0.8410	0.8381	0.8410	0.8378
0.2125	29.76	1860	0.5754	0.8416	0.8393	0.8416	0.8383

Framework versions

Transformers 4.37.2
Pytorch 2.3.0
Datasets 2.19.1
Tokenizers 0.15.1

matthieulel/vit-huge-patch14-224-in21k-finetuned-galaxy10-decals

vit-huge-patch14-224-in21k-finetuned-galaxy10-decals

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

matthieulel/
vit-huge-patch14-224-in21k-finetuned-galaxy10-decals