README.md

metadata

library_name: transformers
license: apache-2.0
base_model: facebook/vit-msn-small
tags:
  - generated_from_trainer
datasets:
  - imagefolder
metrics:
  - accuracy
model-index:
  - name: vit-msn-small-beta-fia-equally-enhanced_test_1
    results:
      - task:
          name: Image Classification
          type: image-classification
        dataset:
          name: imagefolder
          type: imagefolder
          config: default
          split: test
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.8732394366197183

vit-msn-small-beta-fia-equally-enhanced_test_1

This model is a fine-tuned version of facebook/vit-msn-small on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.6061
• Accuracy: 0.8732

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: 1e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.2
• num_epochs: 100
• label_smoothing_factor: 0.1

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.5714	1	1.5578	0.0704
No log	1.7143	3	1.4950	0.0634
No log	2.8571	5	1.3574	0.0634
No log	4.0	7	1.1698	0.1268
No log	4.5714	8	1.0682	0.3169
1.5036	5.7143	10	0.8754	0.7958
1.5036	6.8571	12	0.7359	0.8239
1.5036	8.0	14	0.6782	0.8169
1.5036	8.5714	15	0.6718	0.8169
1.5036	9.7143	17	0.6821	0.8099
1.5036	10.8571	19	0.7157	0.8028
0.7486	12.0	21	0.7173	0.8099
0.7486	12.5714	22	0.6967	0.8169
0.7486	13.7143	24	0.6847	0.8169
0.7486	14.8571	26	0.6827	0.8239
0.7486	16.0	28	0.6959	0.8380
0.7486	16.5714	29	0.6826	0.8521
0.6547	17.7143	31	0.6360	0.8310
0.6547	18.8571	33	0.6257	0.8521
0.6547	20.0	35	0.6594	0.8732
0.6547	20.5714	36	0.6784	0.8380
0.6547	21.7143	38	0.6578	0.8521
0.5817	22.8571	40	0.6146	0.8592
0.5817	24.0	42	0.6212	0.8732
0.5817	24.5714	43	0.6395	0.8732
0.5817	25.7143	45	0.6452	0.8732
0.5817	26.8571	47	0.6317	0.8803
0.5817	28.0	49	0.6332	0.8803
0.5632	28.5714	50	0.6418	0.8732
0.5632	29.7143	52	0.6383	0.8803
0.5632	30.8571	54	0.6367	0.8592
0.5632	32.0	56	0.6253	0.8732
0.5632	32.5714	57	0.6268	0.8592
0.5632	33.7143	59	0.6234	0.8662
0.5328	34.8571	61	0.6368	0.8521
0.5328	36.0	63	0.6251	0.8592
0.5328	36.5714	64	0.6184	0.8732
0.5328	37.7143	66	0.6067	0.8732
0.5328	38.8571	68	0.6182	0.8662
0.5272	40.0	70	0.6398	0.8451
0.5272	40.5714	71	0.6440	0.8310
0.5272	41.7143	73	0.6318	0.8451
0.5272	42.8571	75	0.6111	0.8732
0.5272	44.0	77	0.6061	0.8732
0.5272	44.5714	78	0.6116	0.8732
0.5255	45.7143	80	0.6320	0.8451
0.5255	46.8571	82	0.6394	0.8310
0.5255	48.0	84	0.6379	0.8310
0.5255	48.5714	85	0.6363	0.8310
0.5255	49.7143	87	0.6282	0.8521
0.5255	50.8571	89	0.6214	0.8592
0.52	52.0	91	0.6195	0.8592
0.52	52.5714	92	0.6170	0.8662
0.52	53.7143	94	0.6169	0.8592
0.52	54.8571	96	0.6174	0.8592
0.52	56.0	98	0.6187	0.8592
0.52	56.5714	99	0.6193	0.8592
0.504	57.1429	100	0.6194	0.8592

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.2.0
• Tokenizers 0.19.1