|
--- |
|
license: apache-2.0 |
|
base_model: google/vit-base-patch16-224-in21k |
|
tags: |
|
- generated_from_trainer |
|
datasets: |
|
- imagefolder |
|
metrics: |
|
- accuracy |
|
- precision |
|
- recall |
|
- f1 |
|
model-index: |
|
- name: emotion_classification_v1.2 |
|
results: |
|
- task: |
|
name: Image Classification |
|
type: image-classification |
|
dataset: |
|
name: imagefolder |
|
type: imagefolder |
|
config: default |
|
split: train[:5000] |
|
args: default |
|
metrics: |
|
- name: Accuracy |
|
type: accuracy |
|
value: 0.625 |
|
- name: Precision |
|
type: precision |
|
value: 0.620708259363687 |
|
- name: Recall |
|
type: recall |
|
value: 0.625 |
|
- name: F1 |
|
type: f1 |
|
value: 0.6034583857987293 |
|
--- |
|
|
|
<!-- 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. --> |
|
|
|
# emotion_classification_v1.2 |
|
|
|
This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the imagefolder dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 1.2401 |
|
- Accuracy: 0.625 |
|
- Precision: 0.6207 |
|
- Recall: 0.625 |
|
- F1: 0.6035 |
|
|
|
## Model description |
|
|
|
A slightly more accurate model compared to previous 1.1 version. More information needed |
|
|
|
## Intended uses & limitations |
|
|
|
This model is fined tune solely for face emotion recognition. |
|
|
|
## 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 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 15 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 | |
|
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:| |
|
| No log | 1.0 | 20 | 1.9487 | 0.3312 | 0.3554 | 0.3312 | 0.2830 | |
|
| No log | 2.0 | 40 | 1.6735 | 0.4437 | 0.4238 | 0.4437 | 0.4232 | |
|
| No log | 3.0 | 60 | 1.5359 | 0.4813 | 0.3990 | 0.4813 | 0.4272 | |
|
| No log | 4.0 | 80 | 1.4249 | 0.5 | 0.4178 | 0.5 | 0.4443 | |
|
| No log | 5.0 | 100 | 1.3733 | 0.5062 | 0.4753 | 0.5062 | 0.4653 | |
|
| No log | 6.0 | 120 | 1.3513 | 0.5188 | 0.5076 | 0.5188 | 0.4908 | |
|
| No log | 7.0 | 140 | 1.2377 | 0.6125 | 0.6163 | 0.6125 | 0.5976 | |
|
| No log | 8.0 | 160 | 1.2354 | 0.6062 | 0.6131 | 0.6062 | 0.5961 | |
|
| No log | 9.0 | 180 | 1.2574 | 0.575 | 0.5847 | 0.575 | 0.5728 | |
|
| No log | 10.0 | 200 | 1.2493 | 0.5813 | 0.5912 | 0.5813 | 0.5776 | |
|
| No log | 11.0 | 220 | 1.1954 | 0.5813 | 0.5795 | 0.5813 | 0.5730 | |
|
| No log | 12.0 | 240 | 1.2283 | 0.5625 | 0.5651 | 0.5625 | 0.5598 | |
|
| No log | 13.0 | 260 | 1.1984 | 0.5625 | 0.5800 | 0.5625 | 0.5643 | |
|
| No log | 14.0 | 280 | 1.2308 | 0.5437 | 0.5523 | 0.5437 | 0.5414 | |
|
| No log | 15.0 | 300 | 1.1665 | 0.5938 | 0.6005 | 0.5938 | 0.5935 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.41.2 |
|
- Pytorch 2.3.0 |
|
- Datasets 2.19.1 |
|
- Tokenizers 0.19.1 |
|
|
