Repository navigation
Add a zoom (strong Wolfe) line search - #169
Merged
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Added
search_along, which searches for a step size along a direction inside the compiled graph, andzoom_line_search, a strong Wolfe search ported from optax'sscale_by_zoom_linesearch. The search is oneLineSearchOpnode whose inner graph is ascanwithuntil, so it stops at the first trial it accepts. On the default and numba backends it matches optax 0.2.8 exactly on six pinned cases, and agrees withscipy.optimize.line_searchwhere the two visit the same trials.JAX and MLX cannot run a
scanthat stops early, so they read the search's namedOpFromGraphpieces from the graph their linker has already rewritten. JAX runs thesteppiece in alax.while_loop. MLX runs allmax_stepstrials and holds the state once one is accepted, becausemx.compilecannot read the stop flag back while it traces. At float32 on the GPU, MLX runs each trial's bookkeeping as one Metal kernel between loss evaluations.Nothing calls the search yet. Wiring it into
lbfgsis the next step toward #58.📚 Documentation preview 📚: https://pytensor-ml--169.org.readthedocs.build/en/169/