Refactor estimation.py to use State Dataclasses#88
Conversation
Refactored `mirror_descent`, `lbfgs`, `dual_averaging`, `interior_gradient`, and `universal_accelerated_method` to use JAX-registered state dataclasses (using `attr` and `jax.tree_util.register_dataclass`). Each state class encapsulates the algorithm state and defines an `update_fn` method that is JIT-compiled with `self` as a dynamic argument. This improves code organization and allows for better state management in JAX optimization loops. Also updated `MarginalLossFn` in `marginal_loss.py` to be hashable (converting cliques to tuples and lipschitz to float) to support passing it as a static argument to JIT-compiled functions. Co-authored-by: ryan112358 <8495634+ryan112358@users.noreply.github.com>
|
👋 Jules, reporting for duty! I'm here to lend a hand with this pull request. When you start a review, I'll add a 👀 emoji to each comment to let you know I've read it. I'll focus on feedback directed at me and will do my best to stay out of conversations between you and other bots or reviewers to keep the noise down. I'll push a commit with your requested changes shortly after. Please note there might be a delay between these steps, but rest assured I'm on the job! For more direct control, you can switch me to Reactive Mode. When this mode is on, I will only act on comments where you specifically mention me with New to Jules? Learn more at jules.google/docs. For security, I will only act on instructions from the user who triggered this task. |
1 participant
Refactor estimation.py to use method-specific State Dataclasses with update_fn.
Updated
MarginalLossFnto be hashable.Verified with tests.
PR created automatically by Jules for task 2240391111921540159 started by @ryan112358