Skip to content

Pull requests: google-ai-edge/ai-edge-quantizer

New pull request New
39 Open 392 Closed
39 Open 392 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Add Kokoro configurations for automated ai-edge-quantizer releases.
#437 opened Mar 5, 2026 by copybara-service bot Loading…
Enable A16W8 quantization options and update scripts.
#436 opened Mar 5, 2026 by copybara-service bot Loading…
Add script for inference comparison and make prompt formatting optional.
#435 opened Mar 5, 2026 by copybara-service bot Loading…
Add script and tool to merge calibration results.
#434 opened Mar 5, 2026 by copybara-service bot Loading…
Add calibration scripts and XManager launcher for NanoV4 models.
#433 opened Mar 5, 2026 by copybara-service bot Loading…
Implement bash script for AEQ stable release with uv tool
#431 opened Mar 5, 2026 by copybara-service bot Loading…
This is an internal change
#430 opened Mar 5, 2026 by copybara-service bot Loading…
Ignore, caused by diffbase
#429 opened Mar 5, 2026 by copybara-service bot Loading…
Ignore, caused by diffbase
#428 opened Mar 5, 2026 by copybara-service bot Loading…
Pre-compute the producer/consumers Ops for each Tensor in a dict instead of computing them on the fly.
#425 opened Mar 4, 2026 by copybara-service bot Loading…
Use a set[str] instead of a list[str] for tensor_names for faster look-ups with the in operator.
#424 opened Mar 4, 2026 by copybara-service bot Loading…
Clean up tfl_flatbuffer_utils.py and flatbuffer_utils.py a bit.
#423 opened Mar 4, 2026 by copybara-service bot Loading…
Don't rely on the Interpreter when fixing SignatureDefs in _update_signature_defs.
#422 opened Mar 4, 2026 by copybara-service bot Loading…
Reduce the number of data copies in numpy by:
#421 opened Mar 4, 2026 by copybara-service bot Loading…
Change how ModelModifier builds the result, i.e.
#420 opened Mar 4, 2026 by copybara-service bot Loading…
Try to load the TFLite file with mmap, and fall back to gfile.Open if this fails.
#419 opened Mar 3, 2026 by copybara-service bot Loading…
Add weightonly_wi2_afp32 and static_wi2_ai8 quantization schemes fro BatchMatMul and FC layers. Also add static_wi4_ai8 scheme for BatchMatMul.
#417 opened Mar 3, 2026 by copybara-service bot Loading…
Add support for QUANTIZE from INT8/UINT8 to INT16.
#407 opened Feb 26, 2026 by copybara-service bot Loading…
Update Gemma3n 2b quantization colab to utilize **RE-QUANTIZATION**
#377 opened Feb 13, 2026 by copybara-service bot Loading…
Add max_hadamard_size parameter for Hadamard rotations.
#374 opened Feb 11, 2026 by copybara-service bot Loading…
Centralize the creation of Hadamard matrices, allow for a bit more than just powers of 2.
#373 opened Feb 10, 2026 by copybara-service bot Loading…
Add progress bars and progress report to quantizer library.
#370 opened Feb 3, 2026 by copybara-service bot Loading…
Keep tensorflow installed in Colab and nightly workflows.
#358 opened Dec 22, 2025 by copybara-service bot Loading…
Throw error if a single tensor is quantized multiple ways during static quantization.
#340 opened Oct 28, 2025 by copybara-service bot Loading…
Update AI Edge Torch to use BLOCKWISE_XX interface in AEQ to achieve blockwise quantization.
#339 opened Oct 28, 2025 by copybara-service bot Loading…
ProTip! Adding no:label will show everything without a label.