Prepare and release 0.1.0 - #1
Merged
Merged
Conversation
Set the version to 0.1.0 and add the Sonatype Central publishing configuration, mirroring dimwit: credential host, publishTo, licenses, scmInfo, developers and an early-semver version scheme, plus the sbt-sonatype and sbt-pgp plugins. A `sonaUploadCore` alias publishes `core` only, since `examples` depends on the unpublished plotwit. Depend on the released dimwit-core 0.1.0 rather than 0.2-SNAPSHOT. A release cannot depend on a snapshot, and this also lets CI drop the temporary step that cloned dimwit and published it locally. `examples` keeps resolving plotwit from the local ivy repo, so the dimwit dependency scheme is relaxed there to allow plotwit's newer snapshot to win. Document the coordinates in the README install section. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
…tion point traceable VocabularyEmbedder reached past DimWit into `Jax.jnp.take` to look up an embedding row. `slice(Axis[Vocab].at(token))` expresses the same thing: the `AxisAtTensorIndex` overload takes a `Tensor0[Int32]` index and removes the sliced axis, giving `Tensor1[Embedding, V]` directly. The `take` alternative left in a comment could not have worked — it needs a `Tensor1` of indices and replaces an axis rather than removing it. Rename Huber's `threshold` to `transitionPoint`, which is what the error message and the tests already called it, and take it as a `Tensor0[Float32]` so it can be traced and scheduled, matching `Perturbation.thin`. The eager `require` goes with it: a traced scalar cannot be validated host-side, so the constraint is stated in the scaladoc instead, as `Perturbation` already does for its probability. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Reverts the Tensor0 change: a traced scalar cannot be checked host-side, and losing the `require` costs more here than gaining a schedulable transition point. The rename to `transitionPoint` stays — it is what the error message and the tests already called it. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
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
1 participant
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.
Sets the version to 0.1.0 and adds the Sonatype Central publishing configuration,
mirroring dimwit. Switches
coreto the released dimwit-core 0.1.0, which also letsCI drop the temporary step that cloned dimwit and published it locally.
Replaces the raw-JAX lookup in
VocabularyEmbedderwithslice, renames Huber'sthresholdtotransitionPoint, and addsReferenceMultiHeadAttentionSuiteso thebatched
MultiHeadAttentionis checked against the readable per-head implementation.Documents
cnn,normalization,init,loss, the learning-rate schedules, thepositional encodings and
VocabularyEmbedder.🤖 Generated with Claude Code