francois-rozet
commented
1 month ago Hello @adrianjav, thank you for the feature request and PR! I think this is a nice and sensible feature to add to Zuko. I will review the PR soon.
On another note, I passed by your poster at NeurIPS last year and enjoyed your paper! I had no idea your implementation was based on Zuko :sweat_smile: I am grateful that you found Zuko useful for your research, but I have to say I am saddened that you did not mention it in the manuscript or the README.
In fact, I notice that you copy-pasted a large portion of Zuko's code in your repository without any credits, which is explicitly prohibited by our license.
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
I really don't want to pass as a prick, but could you please address this?
adrianjav
Description
I would like to enable support for autoregressive models (that is, MAF, NSF, and NAF) to use a specific adjacency matrix (rather than an ordering), which could be sparser than the triangular matrix obtained from the ordering. In case an adjacency matrix is provided, all layers would use that adjacency matrix for the masking, instead of using random permutations between layers.
The motivation for this feature is that sometimes adding extra structure to the normalizing flows can be really useful in practice. Past year, we published a paper showing the benefits of such an architecture in the context of causal generative models, and our source code actually adapted Zuko to use adjacency matrix for this purpose.
Since the changes required are not many, I thought I could make a pull request with this feature so that we can reuse Zuko for future projects with much more ease (since the version from our repository is based un Zuko v0.2.2).
Implementation
The implementation would follow that from our repository, and it basically consists in adding the adjacency matrix as another parameter of the constructor, and use it if provided, rather than building the matrix from the ordering.
I have a pull-request ready, which I will link to this issue after submitting it.
Alternatives
I did not think of any alternatives, since: i) the functionality is already implemented, the only real barrier is that it was not a parameter of the constructor; and ii) we have a working and tested implementation.