MilesCranmer commited on
Commit
651f56a
1 Parent(s): 49c1f08

Prioritize deepcopy performance improvement

Browse files
Files changed (1) hide show
  1. README.md +1 -2
README.md CHANGED
@@ -147,13 +147,12 @@ pd.DataFrame, Results dataframe, giving complexity, MSE, and equations
147
 
148
  # TODO
149
 
 
150
  - [ ] Consider adding mutation for constant<->variable
151
  - [ ] Use NN to generate weights over all probability distribution conditional on error and existing equation, and train on some randomly-generated equations
152
  - [ ] Performance:
153
  - [ ] Use an enum for functions instead of storing them?
154
  - Current most expensive operations:
155
- - [x] deepcopy() before the mutate, to see whether to accept or not.
156
- - Seems like its necessary right now. But still by far the slowest option.
157
  - [ ] Calculating the loss function - there is duplicate calculations happening.
158
  - [ ] Declaration of the weights array every iteration
159
  - [x] Hyperparameter tune
 
147
 
148
  # TODO
149
 
150
+ - [ ] Write our own tree copy operation; deepcopy() is the slowest operation by far.
151
  - [ ] Consider adding mutation for constant<->variable
152
  - [ ] Use NN to generate weights over all probability distribution conditional on error and existing equation, and train on some randomly-generated equations
153
  - [ ] Performance:
154
  - [ ] Use an enum for functions instead of storing them?
155
  - Current most expensive operations:
 
 
156
  - [ ] Calculating the loss function - there is duplicate calculations happening.
157
  - [ ] Declaration of the weights array every iteration
158
  - [x] Hyperparameter tune