Commits

Branch selector

 develop

User selector

zhupengyang

Datepicker

All time

Commit History

Commits on Jul 20, 2023

• [XPU] fuse cast to conv2d/fc in mixed precision model (PaddlePaddle#54493)

  zhupengyang

  committedJul 20, 2023
  4df0093

Commits on Jul 14, 2023

• fix embedding_with_eltwise_add_xpu (PaddlePaddle#55354)

  zhupengyang

  committedJul 14, 2023
  95aab36

Commits on Jul 6, 2023

• fix lite xpu config (PaddlePaddle#55188)

  zhupengyang

  committedJul 6, 2023
  b564595

Commits on Jun 27, 2023

• delete_assign_op_pass (PaddlePaddle#54887)

  zhupengyang

  committedJun 27, 2023
  813266a

Commits on Jun 26, 2023

• delete_repeated_ops_pass and reshape_unstack_concat_fuse_pass (PaddlePaddle#54846)

  zhupengyang

  committedJun 26, 2023
  05bd4a8

Commits on Jun 16, 2023

• lite xpu api & clone (PaddlePaddle#54670)

  zhupengyang

  committedJun 16, 2023
  08c9008

Commits on Jun 14, 2023

• set xpu context at runtime (PaddlePaddle#54587)

  zhupengyang

  committedJun 14, 2023
  d0d7d01

Commits on Jun 9, 2023

• refine xpu inference api (PaddlePaddle#54342)

  zhupengyang

  committedJun 9, 2023
  b62b384

Commits on May 26, 2023

• refine delete_repeated_ops_pass (PaddlePaddle#54122)

  zhupengyang

  committedMay 26, 2023
  2b54661

Commits on May 23, 2023

• add multi_encoder_xpu_fuse_pass case (PaddlePaddle#54025)

  zhupengyang

  committedMay 23, 2023
  7590c7c

Commits on May 22, 2023

• multi_encoder support adaptive seqlen (PaddlePaddle#53982)

  zhupengyang

  committedMay 22, 2023
  664a275

• [xpu][infer] support runtime configs (PaddlePaddle#53595)

  zhupengyang

  committedMay 22, 2023
  e135069

Commits on May 5, 2023

• fix xpu max size (PaddlePaddle#53505)

  zhupengyang

  committedMay 5, 2023
  74074a8

Commits on Apr 27, 2023

• xpu quant weight only (PaddlePaddle#53306)

  zhupengyang

  committedApr 27, 2023
  1c97aa6

Commits on Apr 24, 2023

• transform cachekv datalayout of fused_multi_transformer_xpu (PaddlePaddle#53144)

  zhupengyang

  committedApr 24, 2023
  bfa5d6b

Commits on Apr 21, 2023

• optimize write_read_array, gather if beam_size=1 (PaddlePaddle#53130)

  zhupengyang

  committedApr 21, 2023
  e8e9d6c

Commits on Apr 14, 2023

• delete cast if lookup_table_v2 support fp16; delete repeated ops (PaddlePaddle#52888)

  zhupengyang

  committedApr 14, 2023
  7aafeb4

Commits on Apr 13, 2023

• delete useless cast, elementwise_mul (PaddlePaddle#52831)

  zhupengyang

  committedApr 13, 2023
  0695fb8

Commits on Apr 12, 2023

• cache scope in while (PaddlePaddle#52628)

  zhupengyang

  committedApr 12, 2023
  8e7c378

Commits on Mar 31, 2023

• fix xpu fp16 lod_reset (PaddlePaddle#52346)

  zhupengyang

  committedMar 31, 2023
  b413733

Commits on Mar 30, 2023

• [XPU] add delete_cast_op_pass (PaddlePaddle#52305)

  zhupengyang

  committedMar 30, 2023
  8b622d5

• [XPU] add delete_concat_op_pass (PaddlePaddle#52304)

  zhupengyang

  committedMar 30, 2023
  70ebef8

Commits on Mar 29, 2023

• [XPU] optimize pass (PaddlePaddle#52099)

  zhupengyang

  committedMar 29, 2023
  599388e

Commits on Mar 22, 2023

• [XPU] optimize graph if beam_size=1 (PaddlePaddle#51732)

  zhupengyang

  committedMar 22, 2023
  720b14e

• [XPU] subgraph support device param copy (PaddlePaddle#51876)

  zhupengyang

  committedMar 22, 2023
  f293e36

Commits on Mar 13, 2023

• fix phi xpu kernel tensor transform (PaddlePaddle#51306)

  zhupengyang

  committedMar 13, 2023
  fcab331

• [xpu] optimize multi_encoder_xpu_fuse_pass performance (PaddlePaddle#51346)

  zhupengyang

  committedMar 13, 2023
  e2cdd4a

Commits on Mar 7, 2023

• [XPU] support shared weight; delete isolated node (PaddlePaddle#51108)

  zhupengyang

  committedMar 7, 2023
  39a9aba

Commits on Mar 1, 2023

• [XPU] delete op device (PaddlePaddle#51029)

  zhupengyang

  committedMar 1, 2023
  c930994

Commits on Feb 28, 2023

• [XPU] support convert fp16 model (PaddlePaddle#50790)

  zhupengyang

  committedFeb 28, 2023
  f265a31

Commits on Feb 23, 2023

• [XPU] optimize multi_encoder_xpu_pass (PaddlePaddle#50759)

  zhupengyang

  committedFeb 23, 2023
  5c9299e

Commits on Feb 22, 2023

• [XPU] link out_max to x_max between xpu_fusion_ops (PaddlePaddle#50690)

  zhupengyang

  committedFeb 22, 2023
  1fd1c16

Commits on Feb 17, 2023

• [XPU] add multi_encoder_xpu_slice_fuse_pass, generate_sequence_xpu_fuse_pass, generate_sequence_xpu kernel (PaddlePaddle#50570)

  zhupengyang

  committedFeb 17, 2023
  61469ee

Commits on Feb 16, 2023

• [XPU] fix dropout pass; add multi_encoder_xpu_fuse_pass & multi_encoder_xpu kernel (PaddlePaddle#50499)

  zhupengyang

  committedFeb 16, 2023
  c8aa640

Commits on Feb 10, 2023

• [XPU] add fc_xpu op&pass to optimize ernie model (PaddlePaddle#50277)

  zhupengyang

  committedFeb 10, 2023
  945f918

Pagination

PreviousNext