631 Commits (9fe843fa398dee90c1fb2aeb24ddc8da184830d2)

Author SHA1 Message Date
  Chen Xin 9fe843fa39
fix wrong param id of Squeeze for modelwriter (#4658) 3 years ago
  nihui f7af84f001
pnnx reset maxpool return_indices if only output found, drop convtransposed output_size arg for static output padding (#4654) 3 years ago
  He Yang 9a5b6ce20e
fix fprintf type mismatch. (#4647) 3 years ago
  nihui 91cc508a23
pnnx recognize foldable scalar and fuse it into expression (#4639) 3 years ago
  nihui 5b5e9ea537
fix some pnnx build warnings (#4634) 3 years ago
  nihui f8e32aba9c
fix pnnx gru rnn with optional output, fix #4608 (#4631) 3 years ago
  nihui 8066c76bc5
pnnx complex data type and torch.stft family (#4627) 3 years ago
  nihui 8049623d31
pnnx convert torch.mm (#4589) 3 years ago
  nihui 693535afc1
pnnx torch 2.0 (#4579) 3 years ago
  nihui ae4f630467
pnnx fuse multiheadattention (#4544) 3 years ago
  nihui c68266efd0
pnnx eliminate reshape shape expression for only one dynamic dimsize (#4548) 3 years ago
  nihui afc9310c62
update new operators for modelwriter (#4540) 3 years ago
  nihui 4fc977d7fd
fold scalar constant, write scalar attribute correctly (#4533) 3 years ago
  nihui 1832da8292
concat 4d (#4528) 3 years ago
  nihui 32e2de015e
slice 4d (#4525) 3 years ago
  nihui fc6ce4a641
copyto operator (#4522) 3 years ago
  nihui 242e775d21
pnnx convert torch log10, pow 2 as square (#4518) 3 years ago
  nihui 246e71c526
implement atan2 (#4516) 3 years ago
  Fangjun Kuang 92e75105c9
Support torch.cumsum (#4505) 3 years ago
  nihui ab4cfbf5b0
enrich ncnn binary broadcast rules (#4513) 3 years ago
  Yoh 529c83b8a5
Fuse pixel unshuffle (#4504) 3 years ago
  nihui ce6b80a16b
pnnx flatten input tuple list (#4498) 3 years ago
  向阳 6af0f3777f
Remove redundant include (#4497) 3 years ago
  li mengyang ff4d05a713
fix typo (#4482) 3 years ago
  Yoh 0fffefd3c9
fix crop bug and eliminate Tensor.clone (#4416) 3 years ago
  inisis 62fc16d157
pnnx readme remove duplicate space (#4428) 3 years ago
  nihui aed05aa851
pnnx fuse more function to module (#4351) 3 years ago
  nihui ec1b07c9fe
pnnx fp16 option for ncnn and onnx weight type (#4350) 3 years ago
  nihui 6967baaccc
pnnx convert torch bitwise left_shift right_shift (#4349) 3 years ago
  nihui eceac35a7f
implement MultiheadAttention kdim vdim (#4347) 3 years ago
  nihui 498ca7341b
squeeze and expanddims 4d (#4346) 3 years ago
  Lry89757 6a47f8d15c
gridsample op support (#4288) 3 years ago
  nihui a2af6369d9
match inplace slice copy pattern, rewrite copy uses (#4338) 3 years ago
  nihui a7e3c62a1b
save foldable constants in file for reducing memory usage (#4337) 3 years ago
  nihui cb88e16fdf
pnnx save onnx zero (#4077) 3 years ago
  WuJinxuan abb28435d6
fix:pnnx-softmax (#4333) 3 years ago
  nihui 92da26be79
pnnx load gpu torchscript and reset device (#4330) 3 years ago
  nihui 5b28c1730e
implement ncnn fold and unfold (#4326) 3 years ago
  nihui a12c24d328
pnnx convert fold unfold (#4325) 3 years ago
  nihui b8d40a960f
pnnx convert nn.Softmax2d (#4324) 3 years ago
  nihui bcf06bd1c0
fold new_full and full_like (#4323) 3 years ago
  nihui 0f9a3bb9dc
fix slice end index, fix fp16 model weight alignment (#4317) 3 years ago
  nihui 6e49fa30dc
groupnorm 1d/2d/4d (#4312) 3 years ago
  nihui fdf129fcd7
pnnx pytorch 1.13 (#4314) 3 years ago
  nihui 8edc03c515
pnnx skip dynamic size evaluation (#4291) 3 years ago
  nihui f770987267
pnnx glu batchindex aware conversion (#4285) 3 years ago
  nihui 777e4ef0a0
fix pnnx softmax/normalize/slice negative axis conversion to ncnn (#4284) 3 years ago
  Fangjun Kuang 5281d51535
implement GLU and pnnx conversion (#4283) 3 years ago
  nihui 77eda4c19f
implement lstm proj_size (#4263) 3 years ago
  LiuYi-Up 902954d737
modified the param axes of expanddims in modelwriter (#4259) 3 years ago