Issues

• pooling2d_cl (io_stream) line buffer synthesized with depth 1 in Vitis HLS 2025.1 → C/RTL co-simulation mismatch for MaxPooling2D 2×2

  bug
  fastmachinelearning/hls4ml#1592 · Filter by author DaiYukichi opened on Oct 19, 2026

• Altera: RecurrentReuseFactor has no effect

  good first issue
  fastmachinelearning/hls4ml#1580 · Filter by author vloncar opened on Oct 13, 2026

• QKeras recurrent layers (QSimpleRNN, QLSTM, QGRU): weight quantizers are not applied, so results differ from QKeras (Keras 2 and Keras 3)

  good first issue
  fastmachinelearning/hls4ml#1579 · Filter by author vloncar opened on Oct 13, 2026

• Consider migrating GitLab CI to GitHub Actions

  enhancement
  fastmachinelearning/hls4ml#1557 · Filter by author peter279k opened on Sep 23, 2026

• hls4ml predict() input type and INT8 models

  enhancementgood first issue
  fastmachinelearning/hls4ml#1530 · Filter by author MateusFauri opened on Aug 23, 2026
  11

• hls4ml HGQ2 issues: softmax LUT mismatch + negative bitwidths

  bug
  fastmachinelearning/hls4ml#1523 · Filter by author m-glowacki opened on Aug 10, 2026

• [Bug] hls4ml 1.3.0: HLS softmax LUT causes argmax flip rate vs float32 despite correct normalization.

  bug
  fastmachinelearning/hls4ml#1512 · Filter by author niteshg97 opened on Jul 31, 2026

• QDepthwiseConv2D produces wrong output — alternating channels zeroed in both io_parallel and io_stream (QKeras + Vivado backend)

  bug
  fastmachinelearning/hls4ml#1471 · Filter by author TagwaM-M opened on May 6, 2026
  1

• SimpleRNN layer is partially parsed by hls4ml (weights are generated), but not instantiated in the generated HLS C++ (myproject.cpp

  bug
  fastmachinelearning/hls4ml#1464 · Filter by author jachoulin304 opened on Apr 11, 2026
  2

• test_softmax.py does not test argmax and latency implementations; latency fails

  bug
  fastmachinelearning/hls4ml#1443 · Filter by author vasdommes opened on Feb 19, 2026
  1

• BramFactor Setup and getting: cannot open shared object file: No such file or directory

  bug
  fastmachinelearning/hls4ml#1406 · Filter by author zsrabbani opened on Dec 4, 2025
  1

• Vitis Backend example (until bitstream generation)

  enhancement
  fastmachinelearning/hls4ml#1387 · Filter by author AnouarITI opened on Sep 22, 2025
  1

• Non-optimal II optimization for Pointwise Conv1D

  bugenhancement
  fastmachinelearning/hls4ml#1223 · Filter by author Duchstf opened on Mar 11, 2025
  3

• User-exposed optimiser skipping

  enhancementgood first issue
  fastmachinelearning/hls4ml#1128 · Filter by author bo3z opened on Nov 14, 2024
  1

• DSP packing Vivado backend

  enhancement
  fastmachinelearning/hls4ml#1096 · Filter by author steltze opened on Oct 28, 2024
  3

• LLM support, attention head and transformer architecture support for hls4ml

  enhancement
  fastmachinelearning/hls4ml#1020 · Filter by author danlg opened on Jun 9, 2024
  2

• SR backend doesn't work with Vitis HLS

  bug
  fastmachinelearning/hls4ml#994 · Filter by author bo3z opened on Apr 10, 2024
  2

• Conv2DTranspose is not supported

  enhancement
  fastmachinelearning/hls4ml#896 · Filter by author AnouarITI opened on Oct 24, 2023
  1

• Add Resource, io_parallel, depth_multiplier>1 and Quartus support for (Q)DepthwiseConv

  enhancement
  fastmachinelearning/hls4ml#859 · Filter by author jmitrevs opened on Aug 29, 2023
  2

• Support for Keras v3 file format

  enhancementgood first issuehelp wanted
  fastmachinelearning/hls4ml#831 · Filter by author vloncar opened on Jul 12, 2023
  2

• Save model weights in quantized fixed point representation instead of floating point

  enhancement
  fastmachinelearning/hls4ml#808 · Filter by author rfforelli opened on Jun 13, 2023

• Adding the InstanceNormalization Layer

  enhancementgood first issuehelp wanted
  fastmachinelearning/hls4ml#717 · Filter by author LorenzoValente3 opened on Feb 16, 2023
  1

• transpose_3d for hls_stream missing

  enhancementgood first issuehelp wanted
  fastmachinelearning/hls4ml#712 · Filter by author AnouarITI opened on Feb 13, 2023
  4

• Split FIFO optimization into multiple optimizers, don't call write explictily

  enhancement
  fastmachinelearning/hls4ml#705 · Filter by author jmitrevs opened on Jan 12, 2023

• BatchNormalization failded conversion if it is untrainable

  enhancement
  fastmachinelearning/hls4ml#630 · Filter by author ChiRuiChen opened on Aug 4, 2022