AkshatSurolia commited on
Commit
7a7f81c
1 Parent(s): a3b4dff

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +33 -1
README.md CHANGED
@@ -1,3 +1,35 @@
1
  ---
2
- license: mit
 
 
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ license: apache-2.0
3
+ tags:
4
+ - image-classification
5
+ datasets:
6
+ - Face-Mask18K
7
  ---
8
+
9
+ # Distilled Data-efficient Image Transformer for Face Mask Detection
10
+
11
+ Distilled data-efficient Image Transformer (DeiT) model pre-trained and fine-tuned on Self Currated Custom Face-Mask18K Dataset (18k images, 2 classes) at resolution 224x224. It was first introduced in the paper Training data-efficient image transformers & distillation through attention by Touvron et al.
12
+
13
+ ## Model description
14
+
15
+ This model is a distilled Vision Transformer (ViT). It uses a distillation token, besides the class token, to effectively learn from a teacher (CNN) during both pre-training and fine-tuning. The distillation token is learned through backpropagation, by interacting with the class ([CLS]) and patch tokens through the self-attention layers.
16
+
17
+ Images are presented to the model as a sequence of fixed-size patches (resolution 16x16), which are linearly embedded.
18
+
19
+ ## Training Metrics
20
+ epoch = 2.0
21
+ total_flos = 2078245655GF
22
+ train_loss = 0.0438
23
+ train_runtime = 1:37:16.87
24
+ train_samples_per_second = 9.887
25
+ train_steps_per_second = 0.309
26
+
27
+ ---
28
+
29
+ ## Evaluation Metrics
30
+ epoch = 2.0
31
+ eval_accuracy = 0.9922
32
+ eval_loss = 0.0271
33
+ eval_runtime = 0:03:17.36
34
+ eval_samples_per_second = 18.22
35
+ eval_steps_per_second = 2.28