trungpq/slac-appearance

appearance

This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6381
• Accuracy: 0.9725
• F1 Macro: 0.9510
• Precision Macro: 0.9484
• Recall Macro: 0.9537
• Total Tf: [1558, 44, 1558, 44]

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: 2e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 212
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro	Precision Macro	Recall Macro	Total Tf
0.2724	1.0	213	0.2590	0.9588	0.9294	0.9118	0.9499	[1536, 66, 1536, 66]
0.2408	2.0	426	0.1592	0.9719	0.9518	0.9338	0.9727	[1557, 45, 1557, 45]
0.0874	3.0	639	0.2457	0.9682	0.9446	0.9320	0.9585	[1551, 51, 1551, 51]
0.0554	4.0	852	0.2417	0.9700	0.9484	0.9320	0.9671	[1554, 48, 1554, 48]
0.0369	5.0	1065	0.4010	0.9738	0.9531	0.9518	0.9544	[1560, 42, 1560, 42]
0.0421	6.0	1278	0.4548	0.9757	0.9564	0.9557	0.9571	[1563, 39, 1563, 39]
0.0449	7.0	1491	0.5434	0.9763	0.9570	0.9613	0.9530	[1564, 38, 1564, 38]
0.0336	8.0	1704	0.6023	0.9732	0.9515	0.9549	0.9481	[1559, 43, 1559, 43]
0.0175	9.0	1917	0.5159	0.9725	0.9511	0.9472	0.9552	[1558, 44, 1558, 44]
0.0061	10.0	2130	0.6394	0.9732	0.9516	0.9537	0.9496	[1559, 43, 1559, 43]
0.0016	11.0	2343	0.4762	0.9732	0.9523	0.9478	0.9570	[1559, 43, 1559, 43]
0.0246	12.0	2556	0.5617	0.9744	0.9544	0.9511	0.9578	[1561, 41, 1561, 41]
0.0101	13.0	2769	0.6450	0.9738	0.9531	0.9518	0.9544	[1560, 42, 1560, 42]
0.0004	14.0	2982	0.6274	0.9725	0.9510	0.9484	0.9537	[1558, 44, 1558, 44]
0.0049	15.0	3195	0.6381	0.9725	0.9510	0.9484	0.9537	[1558, 44, 1558, 44]

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0
• Datasets 2.21.0
• Tokenizers 0.19.1