---
license: mit
base_model: laion/CLIP-ViT-H-14-laion2B-s32B-b79K
tags:
- generated_from_trainer
datasets:
- imagefolder
metrics:
- f1
model-index:
- name: vit-SUPER02
  results:
  - task:
      name: Image Classification
      type: image-classification
    dataset:
      name: imagefolder
      type: imagefolder
      config: default
      split: train
      args: default
    metrics:
    - name: F1
      type: f1
      value: 0.9434741496133529
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# vit-SUPER02

This model is a fine-tuned version of [laion/CLIP-ViT-H-14-laion2B-s32B-b79K](https://huggingface.co/laion/CLIP-ViT-H-14-laion2B-s32B-b79K) on the imagefolder dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2246
- F1: 0.9435

## 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: 64
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | F1     |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 4.2088        | 0.3   | 50   | 4.0561          | 0.0670 |
| 2.9768        | 0.6   | 100  | 2.8147          | 0.2790 |
| 1.9733        | 0.9   | 150  | 1.7144          | 0.5560 |
| 1.1587        | 1.2   | 200  | 1.0544          | 0.7479 |
| 0.7738        | 1.5   | 250  | 0.6839          | 0.8392 |
| 0.5944        | 1.8   | 300  | 0.5771          | 0.8660 |
| 0.3746        | 2.1   | 350  | 0.5237          | 0.8636 |
| 0.4313        | 2.4   | 400  | 0.4649          | 0.8927 |
| 0.3874        | 2.69  | 450  | 0.3890          | 0.9015 |
| 0.346         | 2.99  | 500  | 0.3728          | 0.9072 |
| 0.3123        | 3.29  | 550  | 0.3296          | 0.9113 |
| 0.2976        | 3.59  | 600  | 0.3369          | 0.9166 |
| 0.2371        | 3.89  | 650  | 0.3207          | 0.9139 |
| 0.1462        | 4.19  | 700  | 0.2997          | 0.9195 |
| 0.178         | 4.49  | 750  | 0.2870          | 0.9317 |
| 0.1489        | 4.79  | 800  | 0.3048          | 0.9216 |
| 0.1135        | 5.09  | 850  | 0.2626          | 0.9364 |
| 0.0992        | 5.39  | 900  | 0.2920          | 0.9291 |
| 0.0879        | 5.69  | 950  | 0.2536          | 0.9365 |
| 0.0908        | 5.99  | 1000 | 0.2315          | 0.9435 |
| 0.071         | 6.29  | 1050 | 0.2542          | 0.9378 |
| 0.047         | 6.59  | 1100 | 0.2517          | 0.9426 |
| 0.0565        | 6.89  | 1150 | 0.2513          | 0.9365 |
| 0.0119        | 7.19  | 1200 | 0.2293          | 0.9431 |
| 0.0142        | 7.49  | 1250 | 0.2454          | 0.9414 |
| 0.0124        | 7.78  | 1300 | 0.2391          | 0.9432 |
| 0.0057        | 8.08  | 1350 | 0.2355          | 0.9446 |
| 0.0041        | 8.38  | 1400 | 0.2242          | 0.9520 |
| 0.0107        | 8.68  | 1450 | 0.2230          | 0.9466 |
| 0.009         | 8.98  | 1500 | 0.2236          | 0.9495 |
| 0.0027        | 9.28  | 1550 | 0.2274          | 0.9466 |
| 0.0027        | 9.58  | 1600 | 0.2241          | 0.9454 |
| 0.0022        | 9.88  | 1650 | 0.2246          | 0.9435 |


### Framework versions

- Transformers 4.38.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.1