Augusto777/swiftformer-xs-dmae-va-U5-80C

swiftformer-xs-dmae-va-U5-42C

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: 0.6744
• Accuracy: 0.8167

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: 0.0002
• 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: 80

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.9	7	1.3856	0.4667
1.3855	1.94	15	1.3819	0.4833
1.3855	2.97	23	1.3687	0.4333
1.3742	4.0	31	1.3189	0.3167
1.3004	4.9	38	1.2501	0.4833
1.3004	5.94	46	1.2268	0.4833
1.1716	6.97	54	1.2115	0.5
1.0686	8.0	62	1.2243	0.5333
1.0686	8.9	69	1.1432	0.55
0.9764	9.94	77	1.0205	0.55
0.873	10.97	85	0.9721	0.6
0.873	12.0	93	0.9221	0.5667
0.7822	12.9	100	0.8593	0.6167
0.664	13.94	108	0.7775	0.7
0.664	14.97	116	0.8117	0.6167
0.5439	16.0	124	0.7553	0.6833
0.5439	16.9	131	0.6697	0.7167
0.496	17.94	139	0.6480	0.7333
0.4563	18.97	147	0.7115	0.7333
0.4563	20.0	155	0.6777	0.7167
0.3831	20.9	162	0.6416	0.7667
0.339	21.94	170	0.7040	0.7
0.339	22.97	178	0.6859	0.7167
0.3033	24.0	186	0.6012	0.7
0.2655	24.9	193	0.5440	0.7833
0.2655	25.94	201	0.6174	0.75
0.2269	26.97	209	0.5745	0.7333
0.2472	28.0	217	0.5688	0.8
0.2472	28.9	224	0.6578	0.75
0.2004	29.94	232	0.5811	0.7833
0.2099	30.97	240	0.6672	0.75
0.2099	32.0	248	0.5927	0.75
0.1834	32.9	255	0.6193	0.7667
0.1834	33.94	263	0.7505	0.7167
0.2248	34.97	271	0.7730	0.7333
0.1571	36.0	279	0.6211	0.7333
0.1571	36.9	286	0.6228	0.7167
0.1983	37.94	294	0.6088	0.7167
0.1629	38.97	302	0.7009	0.75
0.1629	40.0	310	0.7285	0.75
0.1547	40.9	317	0.6401	0.7667
0.1548	41.94	325	0.6123	0.7833
0.1548	42.97	333	0.6317	0.8
0.1566	44.0	341	0.7579	0.7167
0.1361	44.9	348	0.6653	0.7167
0.1361	45.94	356	0.7401	0.7333
0.1273	46.97	364	0.8404	0.7333
0.1312	48.0	372	0.8388	0.75
0.1312	48.9	379	0.7823	0.7667
0.1307	49.94	387	0.6980	0.7167
0.1307	50.97	395	0.7589	0.75
0.1061	52.0	403	0.6644	0.7333
0.1186	52.9	410	0.7057	0.7333
0.1186	53.94	418	0.6744	0.8167
0.1108	54.97	426	0.6328	0.7667
0.1014	56.0	434	0.6402	0.7833
0.1014	56.9	441	0.6631	0.75
0.1082	57.94	449	0.7001	0.7333
0.1118	58.97	457	0.7898	0.7333
0.1118	60.0	465	0.7644	0.7333
0.1051	60.9	472	0.7767	0.7333
0.0979	61.94	480	0.7440	0.7333
0.0979	62.97	488	0.6827	0.7333
0.0834	64.0	496	0.7008	0.7333
0.0834	64.9	503	0.7243	0.7167
0.0963	65.94	511	0.7656	0.7333
0.0989	66.97	519	0.7332	0.7333
0.0989	68.0	527	0.7624	0.7333
0.107	68.9	534	0.7292	0.75
0.0987	69.94	542	0.7169	0.7333
0.0987	70.97	550	0.7462	0.7333
0.0956	72.0	558	0.6656	0.75
0.0956	72.26	560	0.6873	0.7333

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2+cu118
• Datasets 2.16.1
• Tokenizers 0.15.0