|
--- |
|
license: apache-2.0 |
|
base_model: facebook/dinov2-small |
|
tags: |
|
- image-classification |
|
- vision |
|
- generated_from_trainer |
|
metrics: |
|
- accuracy |
|
- precision |
|
- recall |
|
- f1 |
|
model-index: |
|
- name: dinov2-small-finetuned-galaxy10-decals |
|
results: [] |
|
--- |
|
|
|
<!-- 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. --> |
|
|
|
# dinov2-small-finetuned-galaxy10-decals |
|
|
|
This model is a fine-tuned version of [facebook/dinov2-small](https://huggingface.co/facebook/dinov2-small) on the matthieulel/galaxy10_decals dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.4852 |
|
- Accuracy: 0.8613 |
|
- Precision: 0.8626 |
|
- Recall: 0.8613 |
|
- F1: 0.8615 |
|
|
|
## 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: 128 |
|
- eval_batch_size: 128 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 4 |
|
- total_train_batch_size: 512 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_ratio: 0.1 |
|
- num_epochs: 30 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 | |
|
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:| |
|
| 1.279 | 0.99 | 31 | 0.9926 | 0.6454 | 0.6589 | 0.6454 | 0.6242 | |
|
| 0.8187 | 1.98 | 62 | 0.6014 | 0.7948 | 0.8151 | 0.7948 | 0.7948 | |
|
| 0.77 | 2.98 | 93 | 0.6323 | 0.7818 | 0.7928 | 0.7818 | 0.7777 | |
|
| 0.6743 | 4.0 | 125 | 0.5555 | 0.8140 | 0.8149 | 0.8140 | 0.8083 | |
|
| 0.6407 | 4.99 | 156 | 0.5732 | 0.8078 | 0.8174 | 0.8078 | 0.8096 | |
|
| 0.64 | 5.98 | 187 | 0.4912 | 0.8360 | 0.8344 | 0.8360 | 0.8321 | |
|
| 0.5685 | 6.98 | 218 | 0.5089 | 0.8185 | 0.8195 | 0.8185 | 0.8163 | |
|
| 0.5438 | 8.0 | 250 | 0.4806 | 0.8320 | 0.8325 | 0.8320 | 0.8293 | |
|
| 0.5455 | 8.99 | 281 | 0.4781 | 0.8410 | 0.8442 | 0.8410 | 0.8406 | |
|
| 0.5002 | 9.98 | 312 | 0.4403 | 0.8478 | 0.8457 | 0.8478 | 0.8452 | |
|
| 0.5003 | 10.98 | 343 | 0.5079 | 0.8151 | 0.8355 | 0.8151 | 0.8156 | |
|
| 0.4828 | 12.0 | 375 | 0.5307 | 0.8191 | 0.8197 | 0.8191 | 0.8147 | |
|
| 0.4674 | 12.99 | 406 | 0.4732 | 0.8444 | 0.8450 | 0.8444 | 0.8431 | |
|
| 0.4646 | 13.98 | 437 | 0.4952 | 0.8281 | 0.8342 | 0.8281 | 0.8251 | |
|
| 0.4477 | 14.98 | 468 | 0.4607 | 0.8523 | 0.8516 | 0.8523 | 0.8514 | |
|
| 0.4224 | 16.0 | 500 | 0.4514 | 0.8506 | 0.8516 | 0.8506 | 0.8495 | |
|
| 0.3751 | 16.99 | 531 | 0.4665 | 0.8484 | 0.8480 | 0.8484 | 0.8472 | |
|
| 0.3874 | 17.98 | 562 | 0.4462 | 0.8489 | 0.8489 | 0.8489 | 0.8477 | |
|
| 0.3675 | 18.98 | 593 | 0.4674 | 0.8484 | 0.8508 | 0.8484 | 0.8471 | |
|
| 0.3434 | 20.0 | 625 | 0.4644 | 0.8512 | 0.8486 | 0.8512 | 0.8462 | |
|
| 0.3332 | 20.99 | 656 | 0.4711 | 0.8557 | 0.8548 | 0.8557 | 0.8525 | |
|
| 0.3187 | 21.98 | 687 | 0.4665 | 0.8534 | 0.8539 | 0.8534 | 0.8524 | |
|
| 0.3039 | 22.98 | 718 | 0.5015 | 0.8439 | 0.8427 | 0.8439 | 0.8423 | |
|
| 0.282 | 24.0 | 750 | 0.4783 | 0.8563 | 0.8563 | 0.8563 | 0.8559 | |
|
| 0.2843 | 24.99 | 781 | 0.5064 | 0.8534 | 0.8519 | 0.8534 | 0.8514 | |
|
| 0.2661 | 25.98 | 812 | 0.5021 | 0.8484 | 0.8476 | 0.8484 | 0.8460 | |
|
| 0.2595 | 26.98 | 843 | 0.4852 | 0.8613 | 0.8626 | 0.8613 | 0.8615 | |
|
| 0.2442 | 28.0 | 875 | 0.4903 | 0.8568 | 0.8546 | 0.8568 | 0.8543 | |
|
| 0.2477 | 28.99 | 906 | 0.4781 | 0.8585 | 0.8569 | 0.8585 | 0.8570 | |
|
| 0.2437 | 29.76 | 930 | 0.4772 | 0.8585 | 0.8580 | 0.8585 | 0.8577 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.37.2 |
|
- Pytorch 2.3.0 |
|
- Datasets 2.19.1 |
|
- Tokenizers 0.15.1 |
|
|
