|
--- |
|
license: apache-2.0 |
|
tags: |
|
- generated_from_trainer |
|
datasets: |
|
- imagefolder |
|
metrics: |
|
- accuracy |
|
- precision |
|
- recall |
|
- f1 |
|
model-index: |
|
- name: swin-base-patch4-window7-224-in22k-finetuned-memes |
|
results: |
|
- task: |
|
name: Image Classification |
|
type: image-classification |
|
dataset: |
|
name: imagefolder |
|
type: imagefolder |
|
config: default |
|
split: train |
|
args: default |
|
metrics: |
|
- name: Accuracy |
|
type: accuracy |
|
value: 0.8562596599690881 |
|
- name: Precision |
|
type: precision |
|
value: 0.8545652818321074 |
|
- name: Recall |
|
type: recall |
|
value: 0.8562596599690881 |
|
- name: F1 |
|
type: f1 |
|
value: 0.8552274649509984 |
|
--- |
|
|
|
<!-- 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. --> |
|
|
|
# swin-base-patch4-window7-224-in22k-finetuned-memes |
|
|
|
This model is a fine-tuned version of [microsoft/swin-base-patch4-window7-224-in22k](https://huggingface.co/microsoft/swin-base-patch4-window7-224-in22k) on the imagefolder dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.7094 |
|
- Accuracy: 0.8563 |
|
- Precision: 0.8546 |
|
- Recall: 0.8563 |
|
- F1: 0.8552 |
|
|
|
## 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: 0.00012 |
|
- 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: 0.1 |
|
- num_epochs: 20 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 | |
|
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:| |
|
| 1.1655 | 0.99 | 20 | 0.8573 | 0.6955 | 0.6953 | 0.6955 | 0.6683 | |
|
| 0.5506 | 1.99 | 40 | 0.5327 | 0.8083 | 0.8050 | 0.8083 | 0.7963 | |
|
| 0.3573 | 2.99 | 60 | 0.4497 | 0.8338 | 0.8339 | 0.8338 | 0.8317 | |
|
| 0.2083 | 3.99 | 80 | 0.4561 | 0.8354 | 0.8450 | 0.8354 | 0.8368 | |
|
| 0.1545 | 4.99 | 100 | 0.4605 | 0.8423 | 0.8458 | 0.8423 | 0.8430 | |
|
| 0.1014 | 5.99 | 120 | 0.4924 | 0.8524 | 0.8571 | 0.8524 | 0.8538 | |
|
| 0.0854 | 6.99 | 140 | 0.5759 | 0.8393 | 0.8452 | 0.8393 | 0.8400 | |
|
| 0.1012 | 7.99 | 160 | 0.5142 | 0.8362 | 0.8378 | 0.8362 | 0.8361 | |
|
| 0.077 | 8.99 | 180 | 0.5647 | 0.8331 | 0.8538 | 0.8331 | 0.8407 | |
|
| 0.0667 | 9.99 | 200 | 0.5294 | 0.8462 | 0.8509 | 0.8462 | 0.8483 | |
|
| 0.0666 | 10.99 | 220 | 0.6038 | 0.8385 | 0.8415 | 0.8385 | 0.8396 | |
|
| 0.0574 | 11.99 | 240 | 0.6384 | 0.8408 | 0.8431 | 0.8408 | 0.8411 | |
|
| 0.0488 | 12.99 | 260 | 0.6305 | 0.8516 | 0.8561 | 0.8516 | 0.8532 | |
|
| 0.0524 | 13.99 | 280 | 0.6411 | 0.8509 | 0.8526 | 0.8509 | 0.8510 | |
|
| 0.0511 | 14.99 | 300 | 0.6462 | 0.8547 | 0.8542 | 0.8547 | 0.8543 | |
|
| 0.0495 | 15.99 | 320 | 0.6869 | 0.8532 | 0.8534 | 0.8532 | 0.8527 | |
|
| 0.0412 | 16.99 | 340 | 0.6643 | 0.8578 | 0.8554 | 0.8578 | 0.8564 | |
|
| 0.0411 | 17.99 | 360 | 0.7214 | 0.8570 | 0.8539 | 0.8570 | 0.8552 | |
|
| 0.0434 | 18.99 | 380 | 0.7037 | 0.8524 | 0.8507 | 0.8524 | 0.8514 | |
|
| 0.0394 | 19.99 | 400 | 0.7094 | 0.8563 | 0.8546 | 0.8563 | 0.8552 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.24.0.dev0 |
|
- Pytorch 1.11.0+cu102 |
|
- Datasets 2.6.1.dev0 |
|
- Tokenizers 0.13.1 |
|
|
