●Change8

v2.18.1

Breaking Changes
📦 tensorflowView on GitHub →
⚠ 2 breaking🐛 1 fixes⚡ 1 deprecations🔧 2 symbols

Summary

This patch updates curl for security, loosens the ml_dtypes version bound, and deprecates tf.lite.Interpreter in favor of ai_edge_litert.interpreter, while dropping TPU support until 2.19.0.

⚠️ Breaking Changes

  • `tf.lite.Interpreter` is deprecated and will be removed; it now redirects to `ai_edge_litert.interpreter`. Update imports to the new location.
  • Tensorflow‑tpu support is omitted in this patch due to sparsecore bugs; upgrade to version 2.19.0 for TPU functionality.

Migration Steps

  1. Replace any usage of `tf.lite.Interpreter` with `ai_edge_litert.interpreter` as described in the migration guide https://ai.google.dev/edge/litert/migration.
  2. If you rely on Tensorflow‑tpu, skip this 2.18.1 patch and upgrade directly to 2.19.0.

🐛 Bug Fixes

  • Loosened the `ml_dtypes` upper bound to `< 1.0.0` to reduce version conflicts.

🔧 Affected Symbols

tf.lite.Interpreterai_edge_litert.interpreter

⚡ Deprecations

  • `tf.lite.Interpreter` is marked for future deletion.