All pull requests

Pull requests

• [Distributed] Key Muon/Mixed sharding 2D buckets by (color, group ranks)

  #79913·DanielSun11 opened Oct 30, 2026Contributor·

  ·
  5

• [Operator Mechanism] Reject non-empty index into a size-0 axis in gather/take_along_axis InferMeta

  contributor
  #79912·lxg5364 opened Oct 30, 2026Contributor·

  ·
  7

• [FSDP] Fix FSDP FT for ZCC and EMA

  #79911·Xing-lil opened Oct 30, 2026Contributor·

  ·
  4

• Diagnose Slice benchmark performance failure

  contributor
  #79910·liym27 with CopilotAI opened Oct 29, 2026·

  ·
  1

  

• Diagnose Slice benchmark failure as timing-sensitive

  contributor
  #79909·liym27 with CopilotAI opened Oct 29, 2026·

  ·
  1

  

• Diagnose Linux-DCU test timeout

  contributor
  #79908·liym27 with CopilotAI opened Oct 29, 2026·

  ·
  3

  

• [Operator Mechanism] Make CPU ReLU match CUDA for NaN and inf values

  contributor
  #79907·kelly-eclipse opened Oct 28, 2026·

  ·

• [FSDP] Support MixedInnerOptimizer in FullyShardFusion

  #79906·zhengshengning opened Oct 28, 2026Contributor·

  ·
  15

• [FSDP] Supp FSDP sync_grads_each_microbatch

  #79903·Xing-lil opened Oct 28, 2026Contributor·

  ·
  9

• [FSDP] Add set_dense_reduce_scatter_streams to spread dense reduce-scatters

  #79902·liym27 opened Oct 28, 2026Contributor·

  ·
  4

• [Bug Fix] Fix NCHW bilinear backward for large upsampling

  #79896·omoYang opened Oct 26, 2026Contributor·

  ·
  3

• [Operator Mechanism] Align interpolate GPU kernel with torch 2.12

  #79895·omoYang opened Oct 26, 2026Contributor·

  ·
  10

• [FSDP] Refactor Paddle FSDP prefetch element-count budget

  #79894·zhengshengning opened Oct 26, 2026Contributor·

  ·
  7

• docs: fix invalid img attribute and broken zh docs link

  contributor
  #79892·Yhm-2007 opened Oct 25, 2026·

  ·
  1

• Fix image tag formatting in README_cn.md

  contributor
  #79891·Yhm-2007 opened Oct 25, 2026·

  ·
  1

• 测试跳过coverage test

  #79890·Ray961123 opened Oct 23, 2026Collaborator·

  ·
  3

• [FSDP] Add reshard_policy and reshard_filter

  #79887·liym27 opened Oct 23, 2026Contributor·

  ·
  6

• [API Compatibility No.144] Add torch-style kwarg aliases for nn.Unflatten -part

  contributor
  #79885·lu123bing opened Oct 21, 2026·

  ·

• [Operator Mechanism] Align embedding kernel with torch2.12

  #79879·omoYang opened Oct 19, 2026Contributor·

  ·
  4

• [Build] Enable FlashAttention v3 with CUDA 13.2

  contributor
  #79869·SigureWang opened Oct 14, 2026Contributor·

  ·
  3

• 修改CE编译选项

  #79865·XieYunshen opened Oct 14, 2026Contributor·

  ·
  5

• [CI] Disable release/3.4 CI

  contributor
  #79864·SigureWang opened Oct 13, 2026Contributor·

  ·

• [Test] add deterministic reducescatter in sharding_optimizer

  #79855·Waynezee opened Oct 10, 2026Contributor·

  ·
  4

• update vmm

  #79849·liuruyan opened Oct 9, 2026Contributor·

  ·
  4

• [CI] Use Paddle-Review-Bot as the only bot gate approver (release/3.4)

  contributor
  #79848·Paddle-Review-Bot opened Oct 9, 2026Contributor·

  ·
  1