prakharg24 commited on
Commit
a0e4212
·
verified ·
1 Parent(s): 297f377

Update my_pages/rashomon_developer.py

Browse files
Files changed (1) hide show
  1. my_pages/rashomon_developer.py +7 -9
my_pages/rashomon_developer.py CHANGED
@@ -12,18 +12,18 @@ def render():
12
 
13
  add_instruction_text(
14
  """
15
- Consider the same data about individuals who did (green) or didn't (red) repay their loans. <br>
16
- Make development choices now that will lead to the same variety of models as before.
17
  """
18
  )
19
 
20
  #### Choosing regularization
21
  st.markdown("""
22
- **Regularization:** Regularization is a technique that prevents the model from “overfitting,” meaning stops them from learning the noise or small quirks in the data, and helps generalize better.
23
 
24
- Choose a regularization technique:
25
- - L1 Regularization: This forces the model to work with as few features as possible, which helps highlight the most important signals and ignore irrelevant ones.
26
- - L2 Regularization: This forces the model to rely less on each feature, even though all features are used, which helps prevent any single feature from overpowering the model.
27
  """
28
  )
29
 
@@ -51,9 +51,7 @@ def render():
51
  #### Choosing random seed
52
  if regularization_method=="l1":
53
  st.markdown("""
54
- **Random Seed:** Random seed controls the stochasticity (or randomness) of the learning process.
55
-
56
- Choose a random seed.
57
  """
58
  )
59
  random_seed = None
 
12
 
13
  add_instruction_text(
14
  """
15
+ Consider the same data as before. <br>
16
+ Instead of directly choosing a model, you make development choices now.
17
  """
18
  )
19
 
20
  #### Choosing regularization
21
  st.markdown("""
22
+ **Regularization:** Regularization is a technique commonly used to stop models from learning the noise or small quirks in the data, which might not generalize.
23
 
24
+ Choose a regularization method:
25
+ - L1 Regularization: Force your AI model to use less number of features, thus avoiding irrelevant features.
26
+ - L2 Regularization: Force your AI model to rely less on each feature, even though you use all features, thus avoiding noisy dominance of any single feature.
27
  """
28
  )
29
 
 
51
  #### Choosing random seed
52
  if regularization_method=="l1":
53
  st.markdown("""
54
+ **Randomness:** Sometimes there is randomness in the learning process. Flip the coin below.
 
 
55
  """
56
  )
57
  random_seed = None