nihui
bd5842174f
pnnx reshape improve compatibility with old ncnn ( #6806 )
...
* onnx flatten to reshape
2026-07-06 19:23:00 +08:00
nihui
81e8cbffd0
pnnx ci update torch 2.12 ( #6796 )
2026-07-03 23:46:56 +08:00
nihui
b10a8327fb
pnnx ncnn legalize global pooling layout ( #6795 )
2026-07-03 16:21:59 +08:00
nihui
127b414614
reshape support batch to/from dim ( #6784 )
2026-07-02 19:29:34 +08:00
nihui
706c4d1964
support 4d mat in more operators ( #6737 )
...
Add 4d Mat handling across generic and optimized backends, extend operator tests, and document supported input and output Mat dimensions.
2026-05-22 15:31:01 +08:00
Lune and nihui
bedf9c6568
unaryop: extend sign, hyperbolic, expm1 and log1p support ( #6675 )
...
Add backend support for new UnaryOp operations across scalar, SIMD and Vulkan paths, including sign, sinh, cosh, asinh, acosh, atanh, expm1 and log1p.
Wire the new ops through onnx2ncnn and pnnx expression conversion, update generated Python emission, and add regression coverage for the new conversion paths.
Fix portability issues for expm1/log1p math declarations, NCNN_SIMPLEMATH builds, ONNX-enabled pnnx linking, and stabilize unary op tests across fp16, bf16 and torch exporter/version differences.
Co-authored-by: nihui <nihuini@tencent.com >
2026-05-21 16:19:58 +08:00
Yeuvoir
9db7174bc6
pnnx: fix Conv2d padding tuple normalization ( #6694 )
...
Normalize 4-element aten pad values from (left, right, top, bottom) to nn.Conv2d padding order (height, width) when fusing Conv2d modules.
Also add regression coverage for asymmetric ZeroPad2d + depthwise Conv2d to keep the output shape and ncnn pad params correct.
2026-05-21 10:17:58 +08:00
476901a510
pnnx print flops and memops ( #5836 )
...
---------
Co-authored-by: luxincn <2391719912@qq.com >
Co-authored-by: SZUwishion <2559916473@qq.com >
Co-authored-by: luxincn <luxincn@users.noreply.github.com >
Co-authored-by: Wu YuXuan <148472521+SZUwishion@users.noreply.github.com >
Co-authored-by: nihui <nihui@users.noreply.github.com >
Co-authored-by: luxincn <118981961+luxincn@users.noreply.github.com >
Co-authored-by: nihui <171016+nihui@users.noreply.github.com >
2026-05-07 19:28:58 +08:00
aa01024557
pnnx: support npy input tensors ( #6700 )
...
Co-authored-by: lovedream-ms <209071654+lovedream-ms@users.noreply.github.com >
Co-authored-by: lancerstadium <57707690+lancerstadium@users.noreply.github.com >
Co-authored-by: AtomAlpaca <66774326+AtomAlpaca@users.noreply.github.com >
2026-05-06 20:52:04 +08:00
nihui
0d29a8d048
update pnnx torch-2.11 ( #6701 )
2026-05-06 20:06:51 +08:00
nihui
3e97a6e61c
update pnnx ci torch 2.10.0 ( #6592 )
...
* Downgrade actions/cache from v5 to v4
2026-03-10 11:49:14 +08:00
nihui
7d357031db
pnnx fuse adjacent permute, eliminate noop permute ( #6467 )
2025-12-26 14:16:34 +08:00
nihui
a7c1fc2499
pnnx fix ncnn torch stack shape out-of-range crash with negative axis ( #6461 )
2025-12-24 10:17:33 +08:00
nihui
232c7ab244
pnnx ncnn convert torch flatten, add test ( #6462 )
...
* Update upload-artifact action version to v5
2025-12-23 18:36:47 +08:00
nihui
559a8e8cb0
rotaryembed layer ( #6407 )
...
* add deepseek_v3 attention test
* fuse non interleaved rotary embed
* fuse t5 layernorm without gamma
* test qwen2 attention
* sdpa pattern ++
2025-11-19 17:28:08 +08:00
nihui
299b6eae5d
fix sdpa attn mask with single channel ( #6402 )
2025-11-12 19:07:11 +08:00
nihui
d73c0d6fd8
pnnx ncnn convert tensor expand to tile, fix missing tile repeats ( #6399 )
2025-11-12 10:56:57 +08:00
nihui
641a3e5be8
ncnn sdpa layer ( #6397 )
...
* fuse gqa
2025-11-11 11:14:04 +08:00
nihui
27ae758da7
pnnx convert interpolate nearest-exact ( #6376 )
2025-10-28 12:52:57 +08:00
nihui
c7a1af0c7e
pnnx convert Tensor.unflatten ( #6374 )
2025-10-27 17:44:00 +08:00
nihui
8903f443aa
pnnx fuse more transformer attention variants ( #6371 )
...
* pnnx fuse more transformer attention variants
* test new transformers
* always eliminate contiguous
* unified view and reshape
* better float param string
2025-10-27 11:41:12 +08:00
nihui
7b207dd3ad
update pnnx ci torch 2.9.0, support onnx external data, enable dynamo exported onnx test ( #6364 )
2025-10-23 10:07:42 +08:00
nihui
cbf4960496
pnnx convert onnx gelu groupnorm rmsnorm ( #6330 )
2025-09-24 17:35:01 +08:00
nihui
22aba560d8
pnnx convert shrink family to ncnn, fix shrink test ( #6329 )
2025-09-24 15:22:02 +08:00
nihui
a6eda68f23
disable onnx2ncnn build ( #6324 )
...
Migrate most of onnx2ncnn functions to pnnx
2025-09-23 19:06:07 +08:00
nihui
9b46272bdf
pnnx match more pad-conv patterns, fix ops out of index crash in fuse_constantlist ( #6260 )
2025-08-19 11:23:56 +08:00
nihui
d68ea6b4d5
update pnnx ci torch 2.8.0 ( #6238 )
2025-08-07 20:33:46 +08:00
nihui and 佰阅
9b91fe5153
implement flip layer and pnnx torch.flip conversion ( #6233 )
...
Co-authored-by: 佰阅 <43716063+Baiyuetribe@users.noreply.github.com >
2025-08-06 14:49:16 +08:00
nihui
a17e31b2f4
pnnx handles conv convtranspose linear weight norm ( #6224 )
2025-08-01 17:21:59 +08:00
nihui
171b9d1bba
use spdx license header, copyright Tencent ( #6152 )
2025-07-08 11:01:29 +08:00
nihui
ff11080db6
pnnx convert onnx MaxPool auto_pad same, torch reshape_as ( #6089 )
2025-05-29 17:16:00 +08:00
nihui
abb7de9a66
pnnx transformers update part2 ( #6072 )
...
* pnnx update batch 1 mask attention, add more clip test
* fuse same qkv input
* skip onnx clip sdpa
* maintain cache
2025-05-26 12:35:58 +08:00
nihui
d9d9cc403d
pnnx torch logical ops test ( #6069 )
2025-05-22 11:44:15 +08:00
nihui
5f056b8fee
pnnx transformers update, part1 ( #6054 )
...
* local cache
* skip transformer test on torch<2.1
* zip pnnx
* bool8 to bool
* initialize xlnet weights
* onnx einsum
* onnx reducelogsumexp
* skip tensor as scalar attribute
* skip longformer
2025-05-20 14:21:11 +08:00
nihui
1d9c78c526
update pnnx ci ( #6048 )
2025-05-12 20:01:42 +08:00
nihui
b8602a6587
pnnx test_inference with torch.no_grad() ( #6028 )
2025-04-28 18:49:12 +08:00
nihui
65b3530be1
fix pnnx ncnn resolve reshape batch index ( #6021 )
...
* handle batch index with dynamic reshape
2025-04-25 17:20:42 +08:00
nihui
76ce92073e
update pnnx ci with torch 2.7 ( #6018 )
2025-04-24 17:36:10 +08:00
nihui
8363040cb4
pnnx ncnn gelu fast mode, fix interp 2d resize ( #5999 )
2025-04-17 18:53:39 +08:00
nihui
ef0b0e631c
interp output size expression ( #5994 )
2025-04-17 14:24:57 +08:00
nihui
39c055d7f2
crop axes starts ends expression ( #5976 )
...
* skip dynamic tensor index
* handle clone oom
2025-04-14 19:43:50 +08:00
nihui
4e34da40f1
pnnx convert onnx reducel1 reducel2, add onnx norm test ( #5988 )
2025-04-13 00:21:11 +08:00
nihui
3b3b7a94a1
pnnx convert onnx globalavgpool, add onnx adaptive avg/max pool test ( #5987 )
2025-04-11 21:59:04 +08:00
nihui
1d95a9cb41
update pnnx ci with torch 2.6 ( #5972 )
2025-04-03 10:34:36 +08:00
nihui
f3651ce5b8
pnnx convert onnx depth2space, add pixelshuffle pixelunshuffle test ( #5967 )
2025-04-01 15:19:13 +08:00
nihui
524001bf54
pnnx convert onnx conv with asymmatric pads ( #5965 )
...
* pnnx convert onnx conv with asymmatric pads
* pad + conv1d and conv3d
* onnx conv auto_pad
2025-04-01 11:01:27 +08:00
nihui
07267f2618
softmax 4d test and vulkan, softmax unified elempack optimization for x86 arm riscv ( #5931 )
2025-03-18 16:34:19 +08:00
nihui
6396a732ef
reshape shape expression, drop reshape permute, test reshape oom ( #5918 )
2025-03-10 15:31:34 +08:00
nihui
78aca5d30b
pnnx fix onnx clip conversion, add onnx clamp test, always reserve onnx split outputs, convert onnx mod ( #5834 )
2024-12-19 16:57:30 +08:00
nihui
4b0d2dea24
pnnx auto inputshape from traced inputs ( #5825 )
2024-12-18 17:29:12 +08:00
nihui