-
Notifications
You must be signed in to change notification settings - Fork 3
Implement NeuralStatistician #96
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
nmheim
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Good stuff, thanks a lot for the PR! I requested some minor changes in the code :)
nmheim
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Great, we are almost there now! In order to create a new release you just need to increase the version number of GenerativeModels in the Project.toml (from 0.2.2 to 0.2.3). then the julia tagbot will automatically create a new release an register it in the julia registry, so that we can install/update the package via ]add Generativemodels/]update :)
src/statistician.jl
Outdated
| function as | ||
| `loss(x) = lossNS(x, NS)` | ||
| `loss(x) = -elbo(NeuralStatistician,x)` |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This could be misunderstood. It looks like the elbo would only get a NeuralStatistician type instead of an actual object. maybe something like this:
model = NeuralStatistician(cdim, instance_enc, enc_c_dist, cond_z_dist, enc_z_dist, dec_dist)
loss(x) = - elbo(model,x)In that case it would also be really great to have short explanations of the variables in there like instance_enc. but if you are pressed with time we can just merge it with the corrections above.
Codecov Report
@@ Coverage Diff @@
## master #96 +/- ##
==========================================
- Coverage 81.81% 74.57% -7.25%
==========================================
Files 5 6 +1
Lines 88 118 +30
==========================================
+ Hits 72 88 +16
- Misses 16 30 +14
Continue to review full report at Codecov.
|
|
I updated the documentation for NeuralStatistician to include an example and rewrote the docs for |
|
Looks great, thank you! :) |
Adding a new generative model working on sets of data. Work in progress now.