swiftformer-xs-dmae-va-U-SF

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

• Loss: 0.6964
• Accuracy: 0.7615

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 40

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.9	7	1.3887	0.3119
1.4383	1.94	15	1.3440	0.4128
1.3956	2.97	23	1.3159	0.3761
1.36	4.0	31	1.2907	0.3853
1.36	4.9	38	1.2488	0.4404
1.2912	5.94	46	1.2129	0.4037
1.2387	6.97	54	1.1734	0.4679
1.1607	8.0	62	1.1436	0.5138
1.1607	8.9	69	1.0991	0.4954
1.1224	9.94	77	1.0479	0.5505
1.0547	10.97	85	0.9993	0.5963
1.0137	12.0	93	0.9860	0.6147
0.9652	12.9	100	0.9698	0.6147
0.9652	13.94	108	0.9519	0.6055
0.9217	14.97	116	0.9242	0.6055
0.9122	16.0	124	0.9062	0.6147
0.8763	16.9	131	0.8873	0.6422
0.8763	17.94	139	0.8477	0.6514
0.8471	18.97	147	0.8427	0.6514
0.8331	20.0	155	0.8257	0.6881
0.8167	20.9	162	0.8025	0.6881
0.8022	21.94	170	0.8011	0.6972
0.8022	22.97	178	0.8078	0.6972
0.7996	24.0	186	0.7920	0.7064
0.7962	24.9	193	0.7604	0.7248
0.7268	25.94	201	0.7597	0.6972
0.7268	26.97	209	0.7462	0.6972
0.7477	28.0	217	0.7316	0.7064
0.7411	28.9	224	0.7275	0.7523
0.7415	29.94	232	0.7210	0.7248
0.7159	30.97	240	0.7271	0.7248
0.7159	32.0	248	0.7005	0.7431
0.7322	32.9	255	0.7012	0.7431
0.7124	33.94	263	0.7052	0.7523
0.7194	34.97	271	0.6964	0.7615
0.7194	36.0	279	0.7007	0.7523
0.6903	36.13	280	0.6987	0.7431

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.1