Closed nmd2k closed 2 years ago
Merging #27 (e0dccc6) into main (aac5e19) will increase coverage by
17.31%
. The diff coverage is91.27%
.
@@ Coverage Diff @@
## main #27 +/- ##
===========================================
+ Coverage 74.25% 91.57% +17.31%
===========================================
Files 6 6
Lines 268 178 -90
Branches 42 0 -42
===========================================
- Hits 199 163 -36
+ Misses 47 15 -32
+ Partials 22 0 -22
Impacted Files | Coverage Δ | |
---|---|---|
src/uetai/logger/comet.py | 88.72% <88.72%> (ø) |
|
src/uetai/logger/__init__.py | 100.00% <100.00%> (ø) |
|
src/uetai/logger/base.py | 100.00% <100.00%> (ø) |
|
src/uetai/utilities/__init__.py | 100.00% <100.00%> (ø) |
|
src/uetai/utilities/imports.py | 100.00% <100.00%> (ø) |
test passed
Re-scope UETAI
With the last version of
uetai
, I tried to integrate the Weights & Biases dashboard into our tools. The main scope was planned to design custom templates that focus on a particular task (e.g: image segmentation, etc) and automate logging everything that we believe is essential in that task.However, we decided on the goals of this library will be able to support researchers manage their experiments with models. Providing the user the ability to store, manage, arrange metadata or experiments with its asset (e.g: models, dataset), this library targets to help users to visualize and debug their data, model with little effort. In this pull request, I integrated a new dashboard provider named "Comet ML", their dashboard have well-written documents and some customizable panel, where I personal have a lot of idea playing with, promising new helpful function and visualization to user.
For more detail of Comet ML, please check: https://www.comet.ml/docs/