I try to reproduce the results from the notebook: Pipeline Scoring Example.ipynb from the github repository.
This notebook evaluate the f1 score of the pipeline lstm_dynamic_threshold on the S-1 timeserie.
When I run the notebook, whithout changing anything, it works whithout raising any exception.
But, he resulting F-1 score is very different from what is expected from the github result.
It seams that I have such an instability issue.
What I DidWhen I run the notebook, the final F-1 score I get is:
But according to github, the results should be:
So there is a huge difference, I don't know from where it could come.
When I rerun the notebook, I get differents results: the f1 score takes values from: 0.08 to 0.3.
To give a little more details:
According to the logs of the training, it seems not to have huge difference with the github source:
own results:
result from orion's gihub:
But the anomalies I detect is a longer sequence, so that explain the precision and the recall I get is worse:
own result:
result from orion's gihub:
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4
