Published January 10, 2023
| Version v0.8.0
Software
Open
Xilinx/brevitas: Release version 0.8.0
Creators
- 1. @AMD
- 2. Zama.ai
- 3. UC San Diego
- 4. @zama-ai
- 5. unpaired.
- 6. @AMD Research Labs
- 7. @Quansight
Description
What's Changed
- Add support for PyTorch 1.11-1.13.1. Brevitas 0.8 supports for PyTorch 1.5.1 to 1.13.1, with 1.10+ suggested.
- Deprecate support for Python 3.6, 3.7+ is now required.
- Add support for export to ONNX QCDQ for <= int8 quantization.
- Extend support for export to ONNX QOps to <= int8 quantization.
- Add experimental support for export to torch QCDQ for <= int32 quantization, as an entry point for future MLIR integration.
- Add support for QuantRNN, QuantLSTM, w/ support for CIFG, bidirectional layers, shared input-hidden gates, shared quantizers, training-time JIT compilation, and partial export support to ONNX (QONNX and QCDQ).
- Extend support for zero-point for both weights and activations quantization.
- New default asymmetric activation quantizer based on percentile rather than min/max.
- Add more built-in quantizers (symmetric per-channel, asymmetric per-channel, symmetric decoupled per-channel).
- Simplify interface for activation calibration.
- Simplify interface for bias correction.
- Initial support for QuantEmbedding.
- Deprecate support for XIR and PyXIR export flows.
- Many bug fixes and minor improvements.
- @fd0r made their first contribution in https://github.com/Xilinx/brevitas/pull/434
- @omarperacha made their first contribution in https://github.com/Xilinx/brevitas/pull/483
Full Changelog: https://github.com/Xilinx/brevitas/compare/v0.7.1...v0.8.0
Files
Xilinx/brevitas-v0.8.0.zip
Files
(2.0 MB)
Name | Size | Download all |
---|---|---|
md5:367c5a817b1922ca9add950eb9ee1347
|
2.0 MB | Preview Download |
Additional details
Related works
- Is supplement to
- https://github.com/Xilinx/brevitas/tree/v0.8.0 (URL)