README.md

metadata

license: apache-2.0
base_model: microsoft/swin-tiny-patch4-window7-224
tags:
  - generated_from_trainer
datasets:
  - imagefolder
metrics:
  - accuracy
model-index:
  - name: swin-tiny-patch4-window7-224-ve-U13-b-80
    results:
      - task:
          name: Image Classification
          type: image-classification
        dataset:
          name: imagefolder
          type: imagefolder
          config: default
          split: validation
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.782608695652174

swin-tiny-patch4-window7-224-ve-U13-b-80

This model is a fine-tuned version of microsoft/swin-tiny-patch4-window7-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.8901
• Accuracy: 0.7826

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: 80

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.92	6	1.3859	0.1304
1.3859	2.0	13	1.3828	0.2826
1.3859	2.92	19	1.3769	0.3261
1.379	4.0	26	1.3566	0.2826
1.3356	4.92	32	1.3162	0.2391
1.3356	6.0	39	1.2090	0.3696
1.2023	6.92	45	1.1409	0.4130
1.0289	8.0	52	1.0442	0.4565
1.0289	8.92	58	0.9696	0.5217
0.9132	10.0	65	1.0133	0.4348
0.7677	10.92	71	1.0144	0.5
0.7677	12.0	78	1.1377	0.3478
0.6988	12.92	84	0.8171	0.5870
0.6051	14.0	91	0.8983	0.6522
0.6051	14.92	97	0.8593	0.6087
0.5175	16.0	104	0.8189	0.6957
0.429	16.92	110	0.6790	0.7174
0.429	18.0	117	0.7074	0.6304
0.4349	18.92	123	0.8890	0.6957
0.3892	20.0	130	0.9798	0.6739
0.3892	20.92	136	0.8814	0.6739
0.3613	22.0	143	0.8840	0.6522
0.3613	22.92	149	0.7662	0.7391
0.342	24.0	156	0.7884	0.7609
0.2762	24.92	162	1.0268	0.6957
0.2762	26.0	169	0.9206	0.7174
0.2759	26.92	175	0.9080	0.6957
0.2559	28.0	182	0.9379	0.6739
0.2559	28.92	188	1.0121	0.6739
0.2455	30.0	195	0.8252	0.7391
0.2125	30.92	201	0.8501	0.7609
0.2125	32.0	208	1.0365	0.6739
0.2204	32.92	214	1.0470	0.7174
0.1785	34.0	221	0.8834	0.7174
0.1785	34.92	227	1.0780	0.7174
0.1912	36.0	234	0.9328	0.7174
0.1518	36.92	240	0.8901	0.7826
0.1518	38.0	247	1.1069	0.6739
0.166	38.92	253	0.9823	0.7174
0.1775	40.0	260	0.9713	0.6957
0.1775	40.92	266	0.9729	0.7174
0.1344	42.0	273	0.9957	0.7174
0.1344	42.92	279	1.0180	0.7391
0.1232	44.0	286	0.9669	0.7826
0.1267	44.92	292	0.9765	0.6957
0.1267	46.0	299	1.0389	0.7391
0.1548	46.92	305	1.0016	0.7174
0.1267	48.0	312	1.1565	0.7391
0.1267	48.92	318	1.1796	0.7174
0.1403	50.0	325	1.2807	0.6957
0.1096	50.92	331	1.2463	0.6739
0.1096	52.0	338	1.1454	0.6957
0.1106	52.92	344	1.1494	0.6957
0.1116	54.0	351	1.1300	0.6957
0.1116	54.92	357	1.2098	0.7174
0.1229	56.0	364	1.0591	0.7174
0.1235	56.92	370	1.1229	0.6957
0.1235	58.0	377	1.2034	0.7174
0.107	58.92	383	1.0628	0.7174
0.107	60.0	390	1.0070	0.7391
0.107	60.92	396	1.1602	0.7174
0.1167	62.0	403	1.0720	0.7174
0.1167	62.92	409	0.9726	0.7391
0.1113	64.0	416	1.0324	0.7174
0.0838	64.92	422	1.1092	0.7174
0.0838	66.0	429	1.1772	0.6957
0.083	66.92	435	1.1195	0.7174
0.0899	68.0	442	1.0681	0.7174
0.0899	68.92	448	1.1160	0.7174
0.0813	70.0	455	1.1435	0.7174
0.0782	70.92	461	1.1464	0.7174
0.0782	72.0	468	1.1539	0.7174
0.1014	72.92	474	1.1585	0.7174
0.0944	73.85	480	1.1592	0.7174

Framework versions

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