mirror of
https://github.com/vee1e/tflite-micro.git
synced 2026-09-02 18:27:52 +00:00
18b9e6f2a8
An upcoming change switches compress() to emitting DECODE-based models. The Python ops resolver registers the DECODE operator unconditionally, so those models load successfully even in a build without compression support. That breaks this test's original approach, which ran a model through compress() and expected the load to fail. Rewrite it to instead inject a raw COMPRESSION_METADATA entry into the flatbuffer via model_editor, directly exercising the HasCompressionMetadata() detection path for legacy-compressed models. Decoupling the test from compress() output lets it verify the legacy-rejection behavior independently of what compress() emits, so it passes both before and after the switch. BUG=part of #3256 |
||
|---|---|---|
| .. | ||
| tests | ||
| tflite_micro | ||
| BUILD | ||
| numpy_cc_targets.bzl | ||
| py_namespace.bzl | ||
| py_rules.bzl | ||
| python_requirements.in | ||
| python_requirements_3_10.txt | ||
| python_requirements_3_11.txt | ||
| python_requirements_3_12.txt | ||
| python_requirements_3_13.txt | ||
| tensorflow_cc_targets.bzl | ||