vit-base-patch16-224-finetuned-galaxy10-decals

This model is a fine-tuned version of google/vit-base-patch16-224 on the matthieulel/galaxy10_decals dataset. It achieves the following results on the evaluation set:

• Loss: 0.5600
• Accuracy: 0.8461
• Precision: 0.8441
• Recall: 0.8461
• F1: 0.8438

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
1.6246	0.99	31	1.3327	0.5705	0.5426	0.5705	0.5446
0.8602	1.98	62	0.7223	0.7599	0.7554	0.7599	0.7530
0.7228	2.98	93	0.6135	0.7937	0.7880	0.7937	0.7871
0.6575	4.0	125	0.5518	0.8061	0.8011	0.8061	0.8003
0.5776	4.99	156	0.5524	0.8134	0.8225	0.8134	0.8148
0.5278	5.98	187	0.5072	0.8315	0.8371	0.8315	0.8291
0.4844	6.98	218	0.4995	0.8399	0.8415	0.8399	0.8393
0.4678	8.0	250	0.4942	0.8269	0.8290	0.8269	0.8246
0.4599	8.99	281	0.5005	0.8326	0.8310	0.8326	0.8294
0.4177	9.98	312	0.5257	0.8168	0.8203	0.8168	0.8165
0.4003	10.98	343	0.4863	0.8337	0.8325	0.8337	0.8317
0.3936	12.0	375	0.4823	0.8343	0.8332	0.8343	0.8327
0.3771	12.99	406	0.5264	0.8275	0.8350	0.8275	0.8282
0.346	13.98	437	0.5195	0.8326	0.8335	0.8326	0.8299
0.3385	14.98	468	0.4956	0.8427	0.8454	0.8427	0.8402
0.3207	16.0	500	0.5112	0.8427	0.8438	0.8427	0.8379
0.2953	16.99	531	0.5106	0.8439	0.8451	0.8439	0.8404
0.2866	17.98	562	0.5286	0.8365	0.8411	0.8365	0.8361
0.2811	18.98	593	0.5227	0.8416	0.8431	0.8416	0.8404
0.2713	20.0	625	0.5359	0.8360	0.8331	0.8360	0.8329
0.2593	20.99	656	0.5321	0.8410	0.8400	0.8410	0.8390
0.2586	21.98	687	0.5413	0.8433	0.8426	0.8433	0.8421
0.2458	22.98	718	0.5550	0.8399	0.8394	0.8399	0.8387
0.2236	24.0	750	0.5589	0.8377	0.8373	0.8377	0.8369
0.2375	24.99	781	0.5548	0.8450	0.8432	0.8450	0.8434
0.2228	25.98	812	0.5726	0.8337	0.8360	0.8337	0.8325
0.2303	26.98	843	0.5630	0.8416	0.8417	0.8416	0.8409
0.2185	28.0	875	0.5580	0.8450	0.8440	0.8450	0.8437
0.2194	28.99	906	0.5600	0.8461	0.8441	0.8461	0.8438
0.2145	29.76	930	0.5530	0.8461	0.8440	0.8461	0.8439

Framework versions

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