All pull requests

Pull requests

• [ROCm] ROCm-latest (ROCm 10.1) build enablement for Paddle

  #59·soumitra-chatterjee-tech opened Aug 28, 2026·

  ·

• [ROCm] AMD Instinct/ROCm overlay for Paddle (on upstream v3.4)

  #58·soumitra-chatterjee-tech opened Aug 28, 2026·

  ·

• [ROCm] Sync ROCm Paddle fork to upstream develop (v3.0 -> v3.4): upstream code

  #57·soumitra-chatterjee-tech opened Aug 28, 2026·

  ·

• [ROCm] Avoid unsupported BF16 llm_int8 registration

  #55·austin1997 opened Jun 23, 2026·

  ·

• [ROCm] Add CINN HIP BF16 codegen support

  #54·austin1997 opened Jun 23, 2026·

  ·

• [ROCm] Enable BF16 top-p sampling kernel

  #53·austin1997 opened Jun 23, 2026·

  ·

• [ROCm] Use native HIP BF16 rounding in phi bfloat16

  #52·austin1997 opened Jun 23, 2026·

  ·

• [ROCm] Check native BF16 support for AMP

  #51·austin1997 opened Jun 23, 2026·

  ·

• [HIP/ROCm] Enable BF16: register layer_norm bfloat16 kernel + guard conv2d fuse passes

  #49·oldzhu opened Apr 23, 2026·

  ·

• [ROCm] Enable BF16 softmax + gate cuDNN-only conv2d_add fuse passes on HIP

  #48·austin1997 opened Apr 22, 2026·

  ·
  2

• add bf16 support for paddleocr-vl

  #46·paperybox opened Apr 13, 2026·

  ·

• Fix data incorrect for reduceSum function

  #43·jichangjichang opened Jan 10, 2019Collaborator·

  ·

• fix memory leackage

  #27·imcarsonliao opened Sep 13, 2018·

  ·

• Review conv2d 1

  #18·dzhwinter opened Jun 21, 2018·

  ·

• Draft of MIOpen optimization2.

  #16·sabreshao opened May 25, 2018Collaborator·

  ·

• Draft of MIOpen optimization.

  #15·sabreshao opened May 25, 2018Collaborator·

  ·

• [Preview PR] Add HIP support for fluid/memory.

  #14·sabreshao opened Apr 16, 2018Collaborator·

  ·