553 Commits (9376ba71c1251f82994de14cee0a662677037c0b)

Author SHA1 Message Date
  nihui c5fed063f5
pnnx fuse expression skip foldable (#3875) 4 years ago
  nihui 7441d8a7c8
eliminate noop cat for single input (#3871) 4 years ago
  nihui e7a664c6e5
convert pnnx torch.index_select and torch.scatter_add (#3842) 4 years ago
  nihui 4c7965781f
add pnnx ncnn pass for chain dict output (#3836) 4 years ago
  nihui 026a04f220
convert torch.norm to ncnn, fix F.normalize vector (#3828) 4 years ago
  nihui 0daad605e0
fix make slice expression with dynamic parameters 4 years ago
  tpoisonooo 6fd801b6d7
feat(src/layer): add vision_transformer benchmark (#3730) 4 years ago
  nihui 2dc1ae45fe
pnnx bitwise and compare op (#3791) 4 years ago
  nihui d476191ff1
pnnx export_onnx function (#3784) 4 years ago
  nihui 0aba8af1d3
pnnx swin transformer (#3783) 4 years ago
  nihui f9c1787de9
implement einsum layer and pnnx conversion (#3768) 4 years ago
  nihui ee6402553c
layernorm for vector and mat along w, pnnx convnext end2end test (#3764) 4 years ago
  nihui 318dc4f5f5
convert Tensor.select to ncnn, fuse multiheadattention unpack, add vit_b_32 end2end test (#3763) 4 years ago
  nihui a2b25c88b0
pnnx convert torch.einsum (#3744) 4 years ago
  841973620 a2e25d3fd3
fix Multiplication result may overflow 'int' before it is converted to 'size_t'. (#3757) 4 years ago
  nihui fff16a025d
pnnx fuse static conv (#3743) 4 years ago
  nihui f9c4e90962
catch more foldable constant in static shape inference mode (#3738) 4 years ago
  nihui 57f3017afd
pnnx torch bmm max min (#3736) 4 years ago
  nihui 88da92f9bf
fix pnnx macos install rpath 4 years ago
  nihui a14acbc6dd
pnnx save ncnn bin with fp16 storage (#3715) 4 years ago
  nihui 06fadd374f fix pnnx hardtanh ncnn conversion 4 years ago
  nihui a03c135319 enable policy cmp0091 before project pnnx 4 years ago
  nihui 0f2f25f0af pnnx static msvc crt option 4 years ago
  nihui 462b80052f define NOMINMAX for pnnx windows build 4 years ago
  nihui 0d83ad99f8 link pnnx with pthread, fix minmax issue on windows build 4 years ago
  nihui d42e048b56
pnnx convert torch.addmm (#3634) 4 years ago
  nihui f9663d7726
pnnx support torch 1.11.0 (#3617) 4 years ago
  nihui 41cd40ac1f
convert pnnx torch.unbind and torch.ones/ones_like family (#3583) 4 years ago
  nihui 2880eff264
deconv1d deconv3d (#3584) 4 years ago
  nihuini bc188ece58
update modelwriter for new operators 4 years ago
  nihui 6b2495cc24
add reshape before and after pooling 123d with no batch dimension (#3566) 4 years ago
  nihui 76e32e9ee6
fix interp nearest by scale factor, fix issue #3555 (#3565) 4 years ago
  nihuini 6ee5ab72f6
handle reshape 5d with batch index 0 4 years ago
  nihui 68d0c8bccb apply code-format changes 4 years ago
  nihui fb5d5c011a convert F.pad with negative pads, eliminate noop expression 4 years ago
  nihui d452eca28f
convert torch.matmul, eliminate noop pad and identity op, fuse transpose matmul, fuse select to unbind (#3554) 4 years ago
  nihui b352221ba3 handle pnnx half weight when converting to ncnn 4 years ago
  nihui 3181616439 treat old reduction axes param as failure 4 years ago
  nihui 340b4e673e
pnnx fold constant (#3521) 4 years ago
  nihuini 1be043aad5
convert torch mean/sum/prod reduction with no args 4 years ago
  nihuini b4a755495c
convert pnnx zeros roll remainder 4 years ago
  nihuini 4ba1eb6d2f
assign unique names for all pnnx operator and operand names. fix #3493 4 years ago
  nihuini 457f7d1c63
fix use-after-free, fix #3492 4 years ago
  nihui c5d7f963b9
layer tile (#3491) 4 years ago
  Xiaoyang Chen 4d31c46532
[pnnx] Update README.md (#3487) 4 years ago
  nihui 7d3503c06a
pnnx Tensor index (#3483) 4 years ago
  nihuini 1db16ce9fc
pnnx torch norm stack test 4 years ago
  nihuini 23d3340017
pnnx norm stack 4 years ago
  nihuini e33bdd16e8
pnnx fuse conv1d-bn convtranspose1d-bn 4 years ago
  nihuini f8ca1e7585
fix pnnx crash on unsupported expression 4 years ago