Extract metric from a test for benchmarkAI #74
Conversation
| self.pattern_list.append('time_step: %s\t') | ||
| self.pattern_list.append('speed: %s\t') | ||
|
|
||
| if 'loss' in history_callback.history: |
There was a problem hiding this comment.
please move the metric constants to constant.py file & iterate over those to increase modularity
| else: | ||
| raise ValueError('Incorrect metric parameter') | ||
|
|
||
| def reformat(self, var): |
There was a problem hiding this comment.
please make this a private method & limit its scope or move it to utils such that LoggingMetrics class can take advantage of it while appending to pattern list
|
As discussed offline, these custom BenchmarkAI specific changes should not be made here in Keras repo. |
|
@karan6181 - Can you please address Kalyanee comments? |
|
@sandeep-krishnamurthy I have already addressed that comments in benchmarkAI repo. Now, this changes is old dated and need some new modifications such as adding runtime epoch parameter, etc. So let me create a new PR and then it's good to go for merge. |
|
This PR code change is old. Created a new PR which has all the latest changes ( #78) |
|
@karan6181 closing this PR as it's out dated |
@sandeep-krishnamurthy @roywei @kalyc