vit-base-patch16-384-finetuned-galaxy10-decals

This model is a fine-tuned version of google/vit-base-patch16-384 on the matthieulel/galaxy10_decals dataset. It achieves the following results on the evaluation set:

• Loss: 0.5422
• Accuracy: 0.8613
• Precision: 0.8600
• Recall: 0.8613
• F1: 0.8596

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: 128
• eval_batch_size: 128
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 512
• 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
1.5894	0.99	31	1.2732	0.5744	0.5409	0.5744	0.5481
0.8001	1.98	62	0.6184	0.7976	0.7934	0.7976	0.7880
0.6895	2.98	93	0.5823	0.8067	0.7991	0.8067	0.7955
0.6259	4.0	125	0.4910	0.8433	0.8427	0.8433	0.8368
0.556	4.99	156	0.4874	0.8467	0.8465	0.8467	0.8465
0.5116	5.98	187	0.4734	0.8546	0.8569	0.8546	0.8518
0.4877	6.98	218	0.4539	0.8461	0.8429	0.8461	0.8428
0.4383	8.0	250	0.4716	0.8377	0.8399	0.8377	0.8345
0.4267	8.99	281	0.4355	0.8602	0.8576	0.8602	0.8559
0.4022	9.98	312	0.4758	0.8377	0.8377	0.8377	0.8356
0.3811	10.98	343	0.4538	0.8495	0.8471	0.8495	0.8474
0.3612	12.0	375	0.4808	0.8439	0.8412	0.8439	0.8399
0.363	12.99	406	0.4751	0.8467	0.8502	0.8467	0.8458
0.3198	13.98	437	0.4800	0.8489	0.8497	0.8489	0.8450
0.3192	14.98	468	0.4834	0.8574	0.8580	0.8574	0.8570
0.3041	16.0	500	0.4879	0.8495	0.8500	0.8495	0.8443
0.2607	16.99	531	0.4958	0.8540	0.8529	0.8540	0.8523
0.2649	17.98	562	0.4927	0.8579	0.8570	0.8579	0.8562
0.2553	18.98	593	0.5095	0.8495	0.8473	0.8495	0.8474
0.2453	20.0	625	0.5162	0.8495	0.8467	0.8495	0.8467
0.2417	20.99	656	0.5375	0.8579	0.8573	0.8579	0.8543
0.241	21.98	687	0.5129	0.8568	0.8546	0.8568	0.8547
0.2257	22.98	718	0.5316	0.8596	0.8584	0.8596	0.8571
0.2087	24.0	750	0.5530	0.8512	0.8497	0.8512	0.8489
0.2196	24.99	781	0.5422	0.8613	0.8600	0.8613	0.8596
0.1975	25.98	812	0.5672	0.8529	0.8534	0.8529	0.8508
0.2135	26.98	843	0.5697	0.8523	0.8513	0.8523	0.8509
0.1946	28.0	875	0.5598	0.8557	0.8542	0.8557	0.8536
0.2006	28.99	906	0.5582	0.8591	0.8566	0.8591	0.8560
0.1968	29.76	930	0.5571	0.8591	0.8571	0.8591	0.8564

Framework versions

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