metadata

license: apache-2.0
base_model: microsoft/swinv2-small-patch4-window8-256
tags:
  - image-classification
  - vision
  - generated_from_trainer
metrics:
  - accuracy
  - precision
  - recall
  - f1
model-index:
  - name: swinv2-small-patch4-window8-256-finetuned-galaxy10-decals
    results: []

swinv2-small-patch4-window8-256-finetuned-galaxy10-decals

This model is a fine-tuned version of microsoft/swinv2-small-patch4-window8-256 on the matthieulel/galaxy10_decals dataset. It achieves the following results on the evaluation set:

Loss: 0.4138
Accuracy: 0.8647
Precision: 0.8650
Recall: 0.8647
F1: 0.8612

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
1.6991	0.99	62	1.4106	0.5011	0.4620	0.5011	0.4641
0.9843	2.0	125	0.8254	0.7148	0.7390	0.7148	0.7097
0.8115	2.99	187	0.6271	0.7773	0.7700	0.7773	0.7671
0.6956	4.0	250	0.5679	0.8061	0.8029	0.8061	0.7967
0.6167	4.99	312	0.5484	0.8281	0.8271	0.8281	0.8247
0.6291	6.0	375	0.5184	0.8191	0.8241	0.8191	0.8123
0.6113	6.99	437	0.5175	0.8134	0.8149	0.8134	0.8097
0.5468	8.0	500	0.4897	0.8309	0.8363	0.8309	0.8283
0.567	8.99	562	0.4459	0.8568	0.8594	0.8568	0.8529
0.5449	10.0	625	0.4544	0.8393	0.8390	0.8393	0.8353
0.5437	10.99	687	0.4528	0.8388	0.8410	0.8388	0.8375
0.4754	12.0	750	0.4524	0.8422	0.8421	0.8422	0.8396
0.5121	12.99	812	0.4840	0.8382	0.8415	0.8382	0.8349
0.5074	14.0	875	0.4138	0.8647	0.8650	0.8647	0.8612
0.4567	14.99	937	0.4339	0.8484	0.8479	0.8484	0.8473
0.4686	16.0	1000	0.4391	0.8540	0.8521	0.8540	0.8504
0.414	16.99	1062	0.4626	0.8422	0.8443	0.8422	0.8388
0.4382	18.0	1125	0.4116	0.8568	0.8558	0.8568	0.8541
0.4322	18.99	1187	0.4506	0.8512	0.8529	0.8512	0.8496
0.4424	20.0	1250	0.4300	0.8568	0.8542	0.8568	0.8538
0.4062	20.99	1312	0.4609	0.8608	0.8597	0.8608	0.8578
0.4459	22.0	1375	0.4517	0.8568	0.8580	0.8568	0.8551
0.4109	22.99	1437	0.4490	0.8534	0.8534	0.8534	0.8513
0.3984	24.0	1500	0.4434	0.8619	0.8606	0.8619	0.8601
0.4034	24.99	1562	0.4613	0.8596	0.8577	0.8596	0.8571
0.3682	26.0	1625	0.4493	0.8596	0.8591	0.8596	0.8573
0.3779	26.99	1687	0.4366	0.8591	0.8581	0.8591	0.8575
0.3965	28.0	1750	0.4370	0.8636	0.8616	0.8636	0.8609
0.3712	28.99	1812	0.4380	0.8591	0.8576	0.8591	0.8568
0.3776	29.76	1860	0.4389	0.8630	0.8624	0.8630	0.8612

Framework versions

Transformers 4.37.2
Pytorch 2.3.0
Datasets 2.19.1
Tokenizers 0.15.1