hls4ml Files

What's Changed

hls4ml v1.0.0 "foxglove" introduces several significant improvements:

• A new QONNX frontend by @jmitrevs introduced in https://github.com/fastmachinelearning/hls4ml/pull/979
• The ability for hls4ml to automatically infer the precision of data types by @vloncar introduced in https://github.com/fastmachinelearning/hls4ml/pull/855
• The addition of an experimental backend for Intel oneAPI by @jmitrevs introduced in https://github.com/fastmachinelearning/hls4ml/pull/955
• The addition of a backend for Siemens Catapult by @dgburnette in https://github.com/fastmachinelearning/hls4ml/pull/956
• Added support for HGQ proxy models by @calad0i in https://github.com/fastmachinelearning/hls4ml/pull/914
• An API for hardware-aware optimization by @bo3z in https://github.com/fastmachinelearning/hls4ml/pull/768 and https://github.com/fastmachinelearning/hls4ml/pull/809

The full list of other improvements and fixes is:

• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in https://github.com/fastmachinelearning/hls4ml/pull/949
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in https://github.com/fastmachinelearning/hls4ml/pull/953
• hls4ml Optimization API [Part 1] by @bo3z in https://github.com/fastmachinelearning/hls4ml/pull/768
• QKeras support for RNN layers by @laurilaatu in https://github.com/fastmachinelearning/hls4ml/pull/856
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in https://github.com/fastmachinelearning/hls4ml/pull/962
• Try to fix sphinx problem by restricting tensorflow-model-optimization by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/967
• Bump pre-commit/action from 3.0.0 to 3.0.1 by @dependabot in https://github.com/fastmachinelearning/hls4ml/pull/968
• Change fractional (and others) to be a property, move quantizers by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/964
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in https://github.com/fastmachinelearning/hls4ml/pull/969
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in https://github.com/fastmachinelearning/hls4ml/pull/971
• vitis backend tarball fix by @calad0i in https://github.com/fastmachinelearning/hls4ml/pull/972
• remove special vitis version of nnet_dense_resource.h by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/975
• Allow Vitis synthesis tests by @jmduarte in https://github.com/fastmachinelearning/hls4ml/pull/927
• Fix cleanup of synthesis tests (leftover from 927) by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/989
• Fix sphinx by pinning tensorflow<=2.15 by @jmduarte in https://github.com/fastmachinelearning/hls4ml/pull/992
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in https://github.com/fastmachinelearning/hls4ml/pull/984
• add clock uncertainty configuration option by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/870
• Stage initial set of changes for the Catapult backend by @dgburnette in https://github.com/fastmachinelearning/hls4ml/pull/956
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in https://github.com/fastmachinelearning/hls4ml/pull/999
• fix unwanted tested file change in [#956] by @calad0i in https://github.com/fastmachinelearning/hls4ml/pull/1000
• Fix SR backend synth missing variables by @bo3z in https://github.com/fastmachinelearning/hls4ml/pull/993
• Upsampling support for PyTorch models by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/977
• Split fpga_types into separate files by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/998
• Support negative_slope in quantized_relu by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/987
• Group more tests per YAML to reduce the number of envs created by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/996
• Automatic precision inference by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/855
• Remove unnecessary transposes related to conversion to channels_last format by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/976
• Update pytest docker image to 0.5.4 by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1005
• Fix pre-commit warning and change '.h5' to '.keras' for written output by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1006
• Fix extension test for Keras v3 by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/1009
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in https://github.com/fastmachinelearning/hls4ml/pull/1007
• updated pytest docker image by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1017
• SepConv1d/2d for io_parallel with Latency strategy by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/1012
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in https://github.com/fastmachinelearning/hls4ml/pull/1021
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in https://github.com/fastmachinelearning/hls4ml/pull/1023
• Latency Pooling Header Updates by @calad0i in https://github.com/fastmachinelearning/hls4ml/pull/973
• Make im2col default option for quartus by @calad0i in https://github.com/fastmachinelearning/hls4ml/pull/1010
• add protection for when kernel_quantizer is None by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/997
• prevent test directory overwrites for activation by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1031
• Update Jenkinsfile to use new Docker image and Python 3.10 environment by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/1033
• clean-up test ci yaml generater by @calad0i in https://github.com/fastmachinelearning/hls4ml/pull/1036
• Add View to layer name map for pytorch parser by @JanFSchulte in https://github.com/fastmachinelearning/hls4ml/pull/1039
• Add RNN support for Pytorch by @JanFSchulte in https://github.com/fastmachinelearning/hls4ml/pull/850
• Add Vitis to pytorch API tests by @JanFSchulte in https://github.com/fastmachinelearning/hls4ml/pull/1040
• clean up mult-dimensional dense by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1042
• Add namespaces and optional writer config by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/986
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in https://github.com/fastmachinelearning/hls4ml/pull/1044
• Add support for HGQ proxy model by @calad0i in https://github.com/fastmachinelearning/hls4ml/pull/914
• Bug Fix for Operand Shape Mismatch in BatchNorm Fusion (PyTorch) by @sei-rquartiano in https://github.com/fastmachinelearning/hls4ml/pull/1045
• remove precision settings that make pytest for batchnorm in pytorch fail by @JanFSchulte in https://github.com/fastmachinelearning/hls4ml/pull/1053
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in https://github.com/fastmachinelearning/hls4ml/pull/1047
• rm slow mnist training in test by @calad0i in https://github.com/fastmachinelearning/hls4ml/pull/1018
• Add an optimizer to replace SeparableConv by Depthwise + Conv (pointwise) by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1022
• Add functionality to use granularity option also for pytorch models by @JanFSchulte in https://github.com/fastmachinelearning/hls4ml/pull/1051
• Update pooling logic for Vivado, Vitis, and Catapult backends by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1056
• remove padding attribute by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1061
• Run long-running pytests out of the batch by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/1062
• Fix tanh activiation in pytorch parser by @JanFSchulte in https://github.com/fastmachinelearning/hls4ml/pull/1055
• make auto the default for layer config by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1016
• remove checks on 'padding' that were missed in previous PR by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1064
• Remove extras flow by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/1067
• Expose alpha and theta type for parametrized activations by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1069
• Raise exception on compile errors by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/1068
• update qkeras in Jenkinsfile by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1072
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in https://github.com/fastmachinelearning/hls4ml/pull/1075
• hls4ml Optimization API [Part 2] by @bo3z in https://github.com/fastmachinelearning/hls4ml/pull/809
• Hardcore weight txt path by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/1089
• quote the ${WEIGHT_DIR} to handle special characters by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1091
• Beginnings of the oneAPI backend by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/955
• update keras activation parsing, especially leaky relu by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1085
• Fix softmax parsing in pytorch and add test by @JanFSchulte in https://github.com/fastmachinelearning/hls4ml/pull/1086
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in https://github.com/fastmachinelearning/hls4ml/pull/1098
• Change indexing in filling result for io_parallel convolutions, Vitis by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1102
• Update QONNX parsing for 1.0 by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/979
• remove incorrect input from Constant nodes by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1119
• add max_precision to onnx parser by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1113
• Add RF to config templates for "Merge" layers by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/1121
• Add doc for HGQ by @calad0i in https://github.com/fastmachinelearning/hls4ml/pull/1117
• Multi output fix 2 by @calad0i in https://github.com/fastmachinelearning/hls4ml/pull/1103
• Make auto default precision for pytorch parser by @JanFSchulte in https://github.com/fastmachinelearning/hls4ml/pull/1112
• remove incorrect setting of result_t by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1130
• Fix problem with scale being a multidimensional array. by @jurevreca12 in https://github.com/fastmachinelearning/hls4ml/pull/1132
• Added support for QONNX Resize node ingestion and tested with tiny UNet model by @nghielme in https://github.com/fastmachinelearning/hls4ml/pull/1122
• Update install_requires for 1.0.0 by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/1136
• Pointwise Conv1D with code generation for "Latency" strategy (update of [#811]) by @jmduarte in https://github.com/fastmachinelearning/hls4ml/pull/881
• Introduce optional description to layer attributes by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/1127
• Qonnx warnings by @jmitrevs in https://github.com/fastmachinelearning/hls4ml/pull/1142
• Fixes to parsing of pytorch models when using torch functionals by @JanFSchulte in https://github.com/fastmachinelearning/hls4ml/pull/1143
• Update README.md for v1.0.0 by @bo3z in https://github.com/fastmachinelearning/hls4ml/pull/1100
• Temporary workaround for QKeras installation by @vloncar in https://github.com/fastmachinelearning/hls4ml/pull/1145

New Contributors

• @laurilaatu made their first contribution in https://github.com/fastmachinelearning/hls4ml/pull/856
• @dgburnette made their first contribution in https://github.com/fastmachinelearning/hls4ml/pull/956
• @sei-rquartiano made their first contribution in https://github.com/fastmachinelearning/hls4ml/pull/1045
• @jurevreca12 made their first contribution in https://github.com/fastmachinelearning/hls4ml/pull/1132

Full Changelog: https://github.com/fastmachinelearning/hls4ml/compare/v0.8.1...v1.0.0