swiftformer-xs-dmae-va-U-40

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: 0.6423
• Accuracy: 0.8165

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	1.3883	0.3211
1.4011	1.94	15	1.3383	0.3578
1.3646	2.97	23	1.2802	0.4404
1.315	4.0	31	1.2194	0.4495
1.315	4.9	38	1.1718	0.5229
1.2634	5.94	46	1.1279	0.5046
1.1949	6.97	54	1.0761	0.5872
1.1136	8.0	62	1.0224	0.6330
1.1136	8.9	69	0.9976	0.6239
1.0824	9.94	77	0.9518	0.6606
1.0212	10.97	85	0.9117	0.6697
0.9566	12.0	93	0.8973	0.6881
0.935	12.9	100	0.8705	0.7064
0.935	13.94	108	0.8559	0.7156
0.8826	14.97	116	0.8371	0.7156
0.8688	16.0	124	0.8252	0.7156
0.8436	16.9	131	0.8211	0.6972
0.8436	17.94	139	0.8040	0.7339
0.8155	18.97	147	0.7625	0.7431
0.7831	20.0	155	0.7452	0.7431
0.7826	20.9	162	0.7279	0.7431
0.7499	21.94	170	0.7148	0.7431
0.7499	22.97	178	0.7061	0.7523
0.7539	24.0	186	0.7026	0.7523
0.7453	24.9	193	0.6819	0.7890
0.7174	25.94	201	0.6837	0.7706
0.7174	26.97	209	0.6743	0.7798
0.7083	28.0	217	0.6706	0.7798
0.6813	28.9	224	0.6644	0.8073
0.7107	29.94	232	0.6423	0.8165
0.6912	30.97	240	0.6419	0.7890
0.6912	32.0	248	0.6465	0.7890
0.7031	32.9	255	0.6346	0.8073
0.6647	33.94	263	0.6347	0.8073
0.6799	34.97	271	0.6476	0.7982
0.6799	36.0	279	0.6429	0.7982
0.6774	36.13	280	0.6518	0.7890

Framework versions

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