Xilinx/brevitas

Issues

• ONNX export of integer weights with large models

  #872 opened 2 months ago by Giuseppe5
  5

• Using external activation functions

  #981 opened 2 months ago by Maya7991
  4

• Feature Request: More padding mode.

  #994 opened 2 months ago by Soptq
  4

• Missing minifloat testing

  #975 opened 2 months ago by Giuseppe5
  0

• Move DelayWrapper logic to Proxy

  #1023 opened 2 months ago by Giuseppe5
  0

• Unsigned quantization of activation function

  #1018 opened 2 months ago by mahdi259
  5

• Modify layers in a pretrained Brevitas Model

  #1008 opened 2 months ago by abedbaltaji
  1

• `ConvXd` -> `QuantConvXd` fails when `padding='same'`

  #1016 opened 2 months ago by nickfraser
  1

• Add tests for TruncAvgPool

  #1015 opened 2 months ago by nickfraser
  0

• Update entrypoint for LLM

  #889 opened 2 months ago by Giuseppe5
  1

• Functionalise Hadamard classifier Module

  #1000 opened 2 months ago by nickfraser
  0

• Replace QuantTensor -> <Datatype>QuantTensor in notebooks

  #999 opened 2 months ago by nickfraser
  0

• Custom bit-width and QuantTensor Operation

  #982 opened 3 months ago by balditommaso
  4

• BrevitasQuantProxyHandler args passing may be sometimes wrong

  #986 opened 3 months ago by andrei-stoian-zama
  9

• Tests: Update QONNX export tests to include property, accuracy validation

  #989 opened 3 months ago by nickfraser
  0

• Assign scale value for QuantIdentity()

  #984 opened 4 months ago by ONQLin
  2

• Docstrings with backslashes are not set as raw strings

  #983 opened 4 months ago by andrei-stoian-zama
  1

• List of all the avalable quantization strategies

  #978 opened 4 months ago by balditommaso
  3

• Weights not quantized after using qnn.QuantConv2d layers for QAT

  #967 opened 4 months ago by Maya7991
  1

• Per-channel zero points but per-tensor scales

  #929 opened 7 months ago by V0XNIHILI
  6

• Mac OSX Tests for `torch==1.9.1` fail when installing dependencies

  #958 opened 5 months ago by nickfraser
  3

• update dependencies=2.0.1 requirement

  #954 opened 5 months ago by isuruf
  4

• Add squeeze / unsqueeze operations to quant invariant functions in `torch_handler.py`

  #891 opened 8 months ago by nickfraser
  4

• Documentation setup thoughts

  #930 opened 7 months ago by V0XNIHILI
  3

• Adding tests for "quantize" function for CNN PTQ

  #908 opened 7 months ago by Giuseppe5
  7

• Call for better/more documentation

  #921 opened 7 months ago by V0XNIHILI
  0

• Missing Proxy tests

  #881 opened 7 months ago by Giuseppe5
  0

• Export ONNX QOperator

  #882 opened 8 months ago by surajpandey353
  5

• Deprecate use of MacOS (Darwin) runners in CI

  #906 opened 8 months ago by Giuseppe5
  0

• Update signature check

  #904 opened 8 months ago by Giuseppe5
  0

• Change way of setting `NaN` and `inf` values for custom minifloat formats

  #901 opened 8 months ago by fabianandresgrob
  0

• Remove maximum assumptions about NaN/inf values for minifloat configurations

  #900 opened 8 months ago by fabianandresgrob
  0

• Implement `torch.where` STE for minifloat clamping

  #899 opened 8 months ago by fabianandresgrob
  0

• Add support for minifloat ptq with fx backend on residual models

  #897 opened 8 months ago by fabianandresgrob
  0

• PTQ Tutorial

  #873 opened 8 months ago by Giuseppe5
  0

• Create GitHub Issue Template

  #855 opened 8 months ago by capnramses
  1

• Test for GPXQ classes

  #854 opened 8 months ago by Giuseppe5
  2

• QuantReLU scale factors

  #859 opened 8 months ago by phixerino
  2

• Fix Value Tracer

  #886 opened 8 months ago by Giuseppe5
  0

• Activation Equalization co-optimize flag

  #887 opened 8 months ago by Giuseppe5
  0

• Question: Unsigned Quantization

  #876 opened 8 months ago by mahdi259
  3

• Create GitHub PR Template

  #856 opened 8 months ago by capnramses
  2

• Implement context-manager based export

  #879 opened 8 months ago by Giuseppe5
  0

• Change how the automatic tests are triggered

  #869 opened 8 months ago by costigt-dev
  1

• Create Brevitas Contributor Doc v1

  #857 opened 8 months ago by capnramses
  3

• Speech to text: Create en empty json file

  #864 opened 8 months ago by rohanjuneja
  2

• Expand QuantAct notebook with scaling per output channel examples

  #862 opened 8 months ago by Giuseppe5
  1

• Notebook Thumbnail in documentation

  #851 opened 8 months ago by Giuseppe5
  0

• QuantTensor always valid

  #844 opened 9 months ago by Giuseppe5
  0

• Explicitly print output of asserts in notebook examples

  #843 opened 9 months ago by nickfraser
  0