-
Our current system only output the best pipeline. They evaluation spec requires us to output all pipelines we searched.
Also, we need rank each pipeline using unique non-negative real numbers.
W…
-
Use python's logging module to log the behavior our ta2 system and ta1 primitives.
Things to log include:
- Primitives executed in dsbox/template/runtime.py
- Metadata annotations add in profil…
-
See these two pages for possible solutions:
https://stackoverflow.com/questions/46911596/why-does-tensorflow-say-a-tensor-is-not-an-element-of-this-graph-when-training-a
https://github.com/tenso…
-
The "ta2-search.py" code does not run in the current #17 . In 'Controller.py':8 we have `import dsbox.template.runtime as runtime` that results in following error:
ModuleNotFoundError: No module n…
-
Add cross-validation capability to Runtime.
Previously we detected and labeled certain primitives `modelers`, then during training phase Runtime.fit() we performed cross-validation on the side. The…
-
- Create Dockerfiles for
- rook services
- the django app
- Create a docker-compose file for local testing
-
In our system the distinction between pipeline and fitted pipeline is not clearly maintained.
A pipeline, i.e. d3m.metadata.pipeline.Pipeline, is just a graph where each node is associated with a p…
-
For some datasets, foreign key is indicated by 'column_index', not 'column_name'
See /dsbox_efs/runs/seed-2018-07-26-02:04/uu2_gp_hyperparameter_estimation/supporting_files/logs/out.txt
```
[IN…
-
https://gitlab.com/datadrivendiscovery/ta3ta2-api/blob/devel/core.proto
checklist: https://docs.google.com/spreadsheets/d/1F00q9yNC3ugyrTW8yl5fcY3ecZQNhod5A-1IyCxAO1U/edit?usp=drive_web&ouid=107433…
-