trpakov commited on
Commit
a2346f6
1 Parent(s): 24d2953

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +46 -0
README.md ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ {}
3
+ ---
4
+ # Vision Transformer (ViT) for Facial Expression Recognition Model Card
5
+
6
+ ## Model Overview
7
+
8
+ - **Model Name:** [trpakov/vit-face-expression](https://huggingface.co/trpakov/vit-face-expression)
9
+
10
+ - **Task:** Facial Expression/Emotion Recognition
11
+
12
+ - **Dataset:** [FER2013](https://www.kaggle.com/datasets/msambare/fer2013)
13
+
14
+ - **Model Architecture:** [Vision Transformer (ViT)](https://huggingface.co/docs/transformers/model_doc/vit)
15
+
16
+ - **Finetuned from model:** [vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k)
17
+
18
+ ## Model Description
19
+
20
+ The vit-face-expression model is a Vision Transformer fine-tuned for the task of facial emotion recognition.
21
+
22
+ It is trained on the FER2013 dataset, which consists of facial images categorized into seven different emotions:
23
+ - Angry
24
+ - Disgust
25
+ - Fear
26
+ - Happy
27
+ - Sad
28
+ - Surprise
29
+ - Neutral
30
+
31
+ ## Data Preprocessing
32
+
33
+ The input images are preprocessed before being fed into the model. The preprocessing steps include:
34
+ - **Resizing:** Images are resized to the specified input size.
35
+ - **Normalization:** Pixel values are normalized to a specific range.
36
+ - **Data Augmentation:** Random transformations such as rotations, flips, and zooms are applied to augment the training dataset.
37
+
38
+ ## Evaluation Metrics
39
+
40
+ - **Validation set accuracy:** 0.7113
41
+ - **Test set accuracy:** 0.7116
42
+
43
+ ## Limitations
44
+
45
+ - **Data Bias:** The model's performance may be influenced by biases present in the training data.
46
+ - **Generalization:** The model's ability to generalize to unseen data is subject to the diversity of the training dataset.