Add optional ts into AbstractPipeline.forecast
#1071
Conversation
This reverts commit d74c777.
|
🚀 Deployed on https://deploy-preview-1071--etna-docs.netlify.app |
Codecov Report
@@ Coverage Diff @@
## inference-v2.1 #1071 +/- ##
===================================================
- Coverage 86.51% 67.93% -18.59%
===================================================
Files 164 164
Lines 8901 8885 -16
===================================================
- Hits 7701 6036 -1665
- Misses 1200 2849 +1649
📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
etna/models/base.py
Outdated
| future_ts.df.loc[:, pd.IndexSlice[segment, feature_nm]] = value[:prediction_size, :] | ||
| # we don't want to change dtype after assignment, but there can happen cast to float32 | ||
| dtype = future_ts.df.loc[:, pd.IndexSlice[segment, feature_nm]] | ||
| future_ts.df.loc[:, pd.IndexSlice[segment, feature_nm]] = value[:prediction_size, :].astype(dtype) |
There was a problem hiding this comment.
May be we should cast to float64 instead if we suppose it
There was a problem hiding this comment.
I just think that using current dtype will be more safe in the future.
|
View / edit / reply to this conversation on ReviewNB martins0n commented on 2023-01-19T07:28:58Z It's interesting why the order has changed 🤔 Mr-Geekman commented on 2023-01-19T07:48:55Z We have non-deterministic order of segments during plot. Do we want to fix that? Mr-Geekman commented on 2023-01-19T07:49:10Z I think it is related to |
|
We have non-deterministic order of segments during plot. Do we want to fix that? View entire conversation on ReviewNB |
|
I think it is related to View entire conversation on ReviewNB |
Before submitting (must do checklist)
Proposed Changes
Look #1069.
Closing issues
Closes #1069.