Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Merge the master branch from tecent/ncnn #6

Merged
merged 86 commits into from
Dec 1, 2022
Merged

Merge the master branch from tecent/ncnn #6

merged 86 commits into from
Dec 1, 2022

Conversation

csukuangfj
Copy link
Owner

Use LSTM with projection from tecent/ncnn so that we can use use_packing_layout = true.

For a test wave, it reduces the decoding time from 14.33 seconds to 1.76 seconds.

MollySophia and others added 30 commits September 4, 2022 14:05
@MollySophia
remove duplicated newline (Tencent#4187)
1d7b217
@Menci
remove duplicated newline (Tencent#4188)
479a73a
@luqiang-guo
optmize softmax arm neon (Tencent#4171)
5148224
@LRY89757
[docs] Fix typo (Tencent#4201)
b16f8ca
@LRY89757
[Prelu x86] Finish intrinsic with elempack merged (Tencent#4177)
9f59711
@magicse
changed size of images for pretty formatting of page (Tencent#4193)
6167b72
@LRY89757
[Gelu x86] Finish intrinsic with elempack merged(fast version) (Tence…
5eb56b2 
…nt#4144)

* Finish the gelu x86 intrinsics
* Finish the fast tanh x86 simd impl
@zchrissirhcz
Ignore .xmake directory (Tencent#4212)
bfe27f2
@dependabot
Bump pypa/cibuildwheel from 2.9.0 to 2.10.1 (Tencent#4207)
183e6e9 
Bumps [pypa/cibuildwheel](https://github.com/pypa/cibuildwheel) from 2.9.0 to 2.10.1.
- [Release notes](https://github.com/pypa/cibuildwheel/releases)
- [Changelog](https://github.com/pypa/cibuildwheel/blob/main/docs/changelog.md)
- [Commits](pypa/cibuildwheel@v2.9.0...v2.10.1)

---
updated-dependencies:
- dependency-name: pypa/cibuildwheel
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <[email protected]>

Signed-off-by: dependabot[bot] <[email protected]>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
@chisaki-takahashi
style: space alignment (Tencent#4217)
d30fc82
@zchrissirhcz
Ignore CMakeSettings.json, the Visual Studio CMake schema file (Tence…
4f9e398 
…nt#4228)
@thelastlin @nihui
RVV: use new interface for segment load/store & change word_type to s…
e7eadca 
…ize_t&add clang ci (part Tencent#4100) (Tencent#4118)

* RVV: use size_t for vl

* RVV: replace vsseg.v tuple type by using regex

-----

search:
vsseg([1-9])e(8|16|32)_v_(f|i|u)\2m(1|2|4|8)x\1\(([ -~]+), vcreate_\3\2m\4x\1\(([ -~]+)\), vl\);

substitute by:
vsseg$1e$2_v_$3$2m$4($5, $6, vl);

* RVV: replace vssseg.v tuple types by using regex

---

search:
vssseg([1-9])e(8|16|32)_v_f\2m1x\1\(([ -~]+), vcreate_f\2m1x\1\(([ -~]+)\), vl\);

substitute by:
vssseg$1e$2_v_f$2m1($3, $4, vl);

* RVV: replace vlseg.v tuple types in load/store

* RVV: replace vloxseg2ei32.v tuple types

* RVV: add a wrapper for old compilers

* RVV: add segment load/store wrapper in pakcing

* RVV: fix cmake test

* RVV: make clang happy by dropping VLAs in sgemm

* RVV: add clang cmake toolchain configure

* RVV: add clang ci, riscv64-unknown-linux-gnu

Co-authored-by: thelastlin <[email protected]>
Co-authored-by: nihui <[email protected]>
@dependabot
Bump pypa/cibuildwheel from 2.10.1 to 2.10.2 (Tencent#4220)
bdcd68f 
Bumps [pypa/cibuildwheel](https://github.com/pypa/cibuildwheel) from 2.10.1 to 2.10.2.
- [Release notes](https://github.com/pypa/cibuildwheel/releases)
- [Changelog](https://github.com/pypa/cibuildwheel/blob/main/docs/changelog.md)
- [Commits](pypa/cibuildwheel@v2.10.1...v2.10.2)

---
updated-dependencies:
- dependency-name: pypa/cibuildwheel
  dependency-type: direct:production
  update-type: version-update:semver-patch
...

Signed-off-by: dependabot[bot] <[email protected]>

Signed-off-by: dependabot[bot] <[email protected]>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
@nihui
add c906 build ci (Tencent#4232)
59a6fa3
@YuzukiTsuru
Add benchmark result of T-Head TH1520 (Tencent#4240)
eb9bb5d 
`cpuinfo`: 

```
isa             : rv64imafdcvsu
mmu             : sv39
cpu-freq                : 1.848Ghz
cpu-icache              : 64KB
cpu-dcache              : 64KB
cpu-l2cache             : 1MB
cpu-tlb         : 1024 4-ways
cpu-cacheline           : 64Bytes
cpu-vector              : 0.7.1
```

Compiled with `-DCMAKE_TOOLCHAIN_FILE=../toolchains/c910-v240.toolchain.cmake -DCMAKE_BUILD_TYPE=release -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF -DNCNN_RUNTIME_CPU=OFF -DNCNN_RVV=ON -DNCNN_SIMPLEOCV=ON -DNCNN_BUILD_EXAMPLES=ON` 

Seems much worse than expected 🤔
@ZhangGe6
fix param parsing issue when layer/blob name exceeds 255 (Tencent#4236)
3fce00b 
* fix param parsing issue when layer/blob name exceeds 255

* apply code-format changes

Co-authored-by: ZhangGe6 <[email protected]>
@LinHeLurking @nihui
Memory Pool Improvement For Variadic Sized Inputs (Tencent#4190)
9426e21 
* Simple miss count for better space efficiency

* Simple double ended greedy;

* Add size drop threshold setter;

* set workspace allocator cr to zero as we had some sort of recylcing capability :P

Co-authored-by: LinHeLurking <[email protected]>
Co-authored-by: nihuini <[email protected]>
@AsabaSeiban
docs: disable fp16 when wrong results encountered caused by overflow (T…
bbbe17c 
…encent#4248)
@nihui
pnnx math operation (Tencent#4251)
cef95f6
@nihui
more stricter armv7 fp16 and armv84 bf16 compiler check, fix Tencent#…
3e2b3fa 
…4147 fix Tencent#4222 (Tencent#4247)
@LiuYi-Up
modified the param axes of expanddims in modelwriter (Tencent#4259)
902954d
@luyanaa
Add TH1520 (4*C910V) toolchain support. (Tencent#4267)
0f38cb2
@nihui
implement lstm proj_size (Tencent#4263)
77eda4c
@miemie2013
Optimize x86 DeformableConv2D (Tencent#4128)
b13c2a1
@veahow
fix compile warning with gcc 9.1.0 including simplestl.h file (Tencen…
f80c274 
…t#4274)

* fix compile warning with gcc 9.1.0 including simplestl.h file

* apply code-format changes

Co-authored-by: veahow <[email protected]>
@hwdef
add benchmark for rk3588 on rock5b (Tencent#4275)
0df463a
@nihui
linux-x64-cpu-gcc on tencent ci
270d6b2
@nihui
implement layer feature disabled bit (Tencent#4278)
0b591b0
@Yoh-Z
add elu vulkan operator (Tencent#4280)
bb660d0
@nihui
fix tencent ci (Tencent#4277)
c62d256
nihui and others added 28 commits November 8, 2022 17:19
@nihui
save foldable constants in file for reducing memory usage (Tencent#4337)
a7e3c62
@nihui
match inplace slice copy pattern, rewrite copy uses (Tencent#4338)
a2af636
@junchao-loongson
add vector optimization for loongarch64 (Tencent#4242)
279222c
@nihui
ci loongarch64 lsx (Tencent#4344)
6019f47
@LRY89757 @nihui
gridsample op support (Tencent#4288)
6a47f8d 

Co-authored-by: LRY89757 <[email protected]>
Co-authored-by: nihuini <[email protected]>
Co-authored-by: nihui <[email protected]>
@nihui
squeeze and expanddims 4d (Tencent#4346)
498ca73
@nihui
implement MultiheadAttention kdim vdim (Tencent#4347)
eceac35
@nihui
pnnx convert torch bitwise left_shift right_shift (Tencent#4349)
6967baa
@nihui
pnnx fp16 option for ncnn and onnx weight type (Tencent#4350)
ec1b07c
@nihui
pnnx fuse more function to module (Tencent#4351)
aed05aa 
* pnnx fuse more function to module

* rename some pass name

* fuse adjacent reshape, fuse pad conv2d

* fuse pad conv1d
@nihui
split tests (Tencent#4354)
057b5bb
@csukuangfj
Support mat.numpy() in Python (Tencent#4356)
1b83fe4
@eltociear
Fix typo in stb_image.h (Tencent#4358)
cdba4ae 
exitting -> exiting
@zchrissirhcz
Fix windows-arm64 build for non-neon case (Tencent#4227)
a5e60ae
@nihui
update release ci (Tencent#4359)
6647396 
* update release ci

* find modern glslang

* parallel jobs on windows
@nihui @qiqikit
Fix c api allocator (Tencent#4360)
0736c5b 
* add some c_api interfaces related to allocator setup.

* fix errors in allocator parameters in c_api.

* test c api allocator

Co-authored-by: zhangtongshe <[email protected]>
@nihui
update glslang (Tencent#4361)
f527fe8
@nihui
disable out-of-line atomics since ndk23+ for resolving linking issue …
cf07bd9 
…with old ndk (Tencent#4362)
@magicse
I added one more project to the list of examples. (Tencent#4205)
8f9a524 
* Dedicated to coloring black and white photographs.
@shaoshengsong
add example project link (Tencent#4365)
47c4ab7
@tpoisonooo
fix(pybind11): build error (Tencent#4368)
bdcbc37
@nihui
fix openmp affinity abort when cpu goes offline (Tencent#4370)
c934c6e
@nihui
Update release-python.yml
03550ba
@csukuangfj
small fixes
77ba238
@csukuangfj
Merge remote-tracking branch 'github/master' into test-master
9e82a82
@csukuangfj
unpack list input
49fd80f
@csukuangfj
Remove LSTM2
064831d
@csukuangfj
fix LSTM
2c6d916
@csukuangfj csukuangfj merged commit 2655d0b into master Dec 1, 2022
@csukuangfj csukuangfj deleted the test-master branch December 1, 2022 04:17
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.