Fix sphinx notebook rendering #57
Conversation
Ohjeah
force-pushed
the
mq/fix-sphinx-notebooks
branch
from
02faa83 to
bfd9d41
Compare
Codecov Report
@@ Coverage Diff @@
## master #57 +/- ##
==========================================
- Coverage 95.52% 95.44% -0.09%
==========================================
Files 18 18
Lines 782 768 -14
==========================================
- Hits 747 733 -14
Misses 35 35
Continue to review full report at Codecov.
|
|
Any idea why the complexity test is failing for just Python 3.8? It looks like In any case, maybe instead of comparing complexity across optimizers, we should look at complexity for a fixed model as the regularization parameter is changed? We could loop over each optimizer and check that as the regularization parameter is increased, the model becomes less complex. Another fix would be to increase the default value of |
|
We could also relax the strict complexity ordering and allow the less complex model to be within +- 1 of the more complex model. |
|
@briandesilva I added the test you suggested and also relaxed the strict condition on my test to account for noise and non-normalized alpha parameters |
…books Fix sphinx notebook rendering
This commit unifies the way the different models preprocess the static and dynamic input data. Closes dynamicslab#57 and dynamicslab#74, part of dynamicslab#64. Similar to the model head, there is now an InputLayer that all (single-frequency) models use. Depending on the config, this layer will: - pass the static variables through an embedding layer (if cfg.statics_embedding is True) - pass the dynamic variables through an embedding layer (if cfg.dynamics_embedding is True) - concatenate static to each step of dynamic inputs (unless this is deactivated; e.g., for EA-LSTM this is not wanted) Key changes: - New config arguments statics_embedding, dynamics_embedding (both bool, default False). - EmbCudaLSTM is no longer needed since the functionality is now part of CudaLSTM. For backwards compatibility, the class still exists but defers to CudaLSTM. - If EA-LSTM and static embeddings are used, there will now be a linear layer between the embedding output and the lstm input to make sure the dimensions match. Previously, the linear layer would be left out if static embeddings were used.
Update to v1.2.0
The current implementation was broken (nothing is currently displayed on sphinx).
I am using a different library now, which has the same look and feel as a regular sphinx gallery.