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)
* pnnx Tensor index
* add test
|
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 |
nihui
|
7c60dc2db7
|
pnnx roialign (#3478)
|
4 years ago |
nihui
|
143258e317
|
pnnx torchvision deformconv2d (#3459)
|
4 years ago |
nihuini
|
de436f9e26
|
pnnx arange matmul zeros_like expand_as
|
4 years ago |
nihui
|
922f8b33c1
|
reduction4d, merge keepdims arg, add test (#3469)
|
4 years ago |
nihuini
|
42017e4f98
|
convert pnnx 4d attribute
|
4 years ago |
nihuini
|
f02b259a15
|
convert some pnnx reduction family
|
4 years ago |
nihuini
|
c5efaf02b7
|
unique attribute identifier
|
4 years ago |
nihuini
|
0532190499
|
convert pnnx expand var
|
4 years ago |
nihuini
|
e0124db194
|
convert pnnx clone normal
|
4 years ago |
nihuini
|
42e7160950
|
eliminate pnnx dropout
|
4 years ago |
nihuini
|
354fc7332d
|
convert pnnx argmax argmin, fix #3435
|
4 years ago |
nihui
|
0533eec14d
|
apply code-format changes
|
4 years ago |
nihuini
|
286e5cef53
|
eliminate pnnx tail reshape permute
|
4 years ago |
nihui
|
6941ec8fc9
|
arm neon optimization for general packed convolution (#3426)
|
4 years ago |
nihui
|
2c56b2f47e
|
apply code-format changes
|
4 years ago |
nihui
|
a3aea3fe99
|
support torch dequantize, quantize_per_tensor and quantized linearrelu, fix #3429
|
4 years ago |
nihui
|
e8046aeb79
|
convert dropout family (#3425)
|
4 years ago |
nihui
|
db1fed6e11
|
pnnx cmdline argument inputshape with type (#3419)
|
4 years ago |
nihui
|
7b0f3fbb78
|
pnnx user provided parameter (#3417)
|
4 years ago |
nihui
|
10ee0eaf5f
|
pnnx write padding_mode zeros for conv
|
4 years ago |
Wang Kangyu
|
1cf229f071
|
[pnnx]: auto copy dll files and remove Torch_INSTALL_DIR option in CI (#3413)
|
4 years ago |
nihui
|
999e640d43
|
dynamic convolution weight (#3408)
|
4 years ago |
Wang Kangyu
|
73c8de6c43
|
auto find pytorch install package (#3394)
|
4 years ago |
nihui
|
f98c396e6b
|
crop4d (#3402)
|
4 years ago |
nihui
|
cf20dbc0bd
|
relu3d, batchnorm3d, reshape4d, flatten4d, permute4d (#3397)
Co-authored-by: ElvisYu <elvisyuovo@gmail.com>
Co-authored-by: 余浩文 <m18107220188@163.com>
Co-authored-by: nihui <nihui@users.noreply.github.com>
Co-authored-by: Zr2223 <67497651+Zr2223@users.noreply.github.com>
Co-authored-by: Zr2223 <Zr2223@users.noreply.github.com>
|
4 years ago |
nihui
|
f10cc6dd93
|
initial data structure changes for 3dcnn, conv3d, pooling3d (#3378)
Co-authored-by: ElvisYu <elvisyuovo@gmail.com>
Co-authored-by: 余浩文 <m18107220188@163.com>
Co-authored-by: Zr2223 <67497651+Zr2223@users.noreply.github.com>
|
4 years ago |
Zhuo Zhang
|
6ef0f51f40
|
fix: do not find opencv when simpleocv specified for examples and quantize tool (#3393)
|
4 years ago |
teng
|
e24a2e6c40
|
[fix] fix typo (#3377)
|
4 years ago |
nihui
|
96d2146926
|
apply code-format changes
|
4 years ago |
nihuini
|
9c742dec01
|
guess batch index from well known operator convention
|
4 years ago |
nihui
|
e4c821a468
|
PNNX is an open standard for PyTorch model interoperability (#3262)
Co-authored-by: nihui <nihui@users.noreply.github.com>
Co-authored-by: xiaooquanwu <546749317@qq.com>
|
4 years ago |
nihuini
|
43cf616000
|
ncnnoptimize embed, fix #3335
|
4 years ago |
YuShiquan
|
0b12f5ee07
|
update readme of quantize (#3283)
|
4 years ago |
zhiliu6
|
814f89ef1a
|
Fuse HardSwish activation into Convolution and InnerProduct (#3233)
* add general fused activation
* add NCNN_FORCE_INLINE option
|
4 years ago |
Jimmy
|
52c6502ae7
|
Feature/ncnntool/add install (#3212)
|
4 years ago |
nihui
|
3d5f447260
|
fix onnx negative concat axis, fix #3194
|
4 years ago |
q.yao
|
41aa601413
|
fix fuse expand (#3189)
|
4 years ago |