feat(compression): implement model_editor for TFLite model manipulation #3259
+1,314
−0
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
rkuester
force-pushed
the
feat-decode-pr
branch
from
b507654 to
054c5ee
Compare
Contributor
Author
|
Rebased to pick up #3261 CI fix. |
rkuester
force-pushed
the
feat-decode-pr
branch
from
054c5ee to
c675328
Compare
Implement unified module for creating, reading, and modifying TFLite models with a clean API. The module eliminates manual index tracking and buffer management through automatic bookkeeping, supporting both declarative and imperative construction styles. Core design uses first-class Buffer objects that can be shared between tensors, with automatic deduplication during build. Tensors reference Buffers directly, matching the TFLite schema structure. The compiler automatically extracts inline tensor declarations, builds operator code tables, and handles index assignment according to TFLite conventions. Supports quantization parameters (per-tensor and per-channel), metadata key-value pairs, and read-modify-write workflows. The read() function preserves the object graph structure, enabling models to be read, modified, and rebuilt. Add comprehensive test coverage for core functionality, advanced features, quantization, and modification workflows.
rkuester
force-pushed
the
feat-decode-pr
branch
from
c675328 to
b17bb64
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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.
Implement unified module for creating, reading, and modifying TFLite
models with a clean API. This provides an easy, readable solution for
creating models in unit tests, unified with our need for editing models
in the compression code itself.
The module eliminates manual index tracking and buffer management
through automatic bookkeeping, supporting both declarative and
imperative construction styles.
Supports quantization parameters (per-tensor and per-channel), metadata
key-value pairs, and read-modify-write workflows.
BUG=part of #3256