ae5df31a28
support layernorm channel first(C#) ( #1204 )
...
* support LayerNorm channel first
* fix review : https://github.com/kendryte/nncase/pull/1204\#discussion_r1599285210
2024-05-14 17:44:31 +08:00
de4b7bd5f8
fix reduce sum opset condition ( #1203 )
...
* fix reduce sum opset condition
* Update Reduce.cs
* fix reduce operators consisting of reduce_sum. fix #1193
2024-05-11 09:32:32 +08:00
482b761f50
qemu system mode ( #1201 )
...
* close non-equal split
* fix typeinfer of conv2d
* fix fold binary
* do not swap binary args
* update setup.py
* do not pre-preprocess onnx with external data
* imgnore .mono in git
* restore SwapBinaryArgs
* fix onnx test runner
---------
Co-authored-by: sunnycase <sunnycase@live.cn>
2024-05-07 12:21:07 +08:00
a2869fa0f8
Feature/add xsgetn for char array buffer ( #1198 )
...
* Load multiple characters at once
* Apply code-format changes
* fix build
---------
Co-authored-by: curioyang <curioyang@users.noreply.github.com>
2024-05-07 10:47:18 +08:00
b17e5b22b6
Fix macos-latest doesn't support python 3.7 issue. ( #1194 )
...
* Fix macos-latest doesn't support python 3.7 issue.
* Set macos version to 12(13 uses Xcode 15).
2024-04-26 17:03:38 +08:00
91ea4df975
fix fold binary ( #1182 )
2024-04-12 12:26:50 +08:00
a13d43d0f4
Feature/update docs ( #1176 )
...
* update faq
* update Homepage
* update link
* update gif link
* update gif link
* update FAQ
* update FAQ
2024-03-13 16:16:47 +08:00
dbcde6f228
Feature/egraph extract constrains ( #1175 )
...
* add egraph-extract-constrains
* reorder SwapBinaryArgs
2024-03-11 16:02:07 +08:00
1cdea27230
Feature/update docs ( #1174 )
...
* update faq
* update Homepage
* update link
* update gif link
* update gif link
2024-03-11 13:58:17 +08:00
bb47ea5803
Revert "add Razor.Templating.Core ( #1169 )" ( #1173 )
...
This reverts commit 2498b1ba0c
2024-03-08 14:40:03 +08:00
2498b1ba0c
add Razor.Templating.Core ( #1169 )
...
* add extract constrains
* refactor buffer schedule
* add Razor.Templating.Core
* reorder SwapBinaryArgs
* Apply code-format changes
* Update NuGet.Config
---------
Co-authored-by: zhengqihang <597323109@qq.com>
Co-authored-by: xhuohai <xhuohai@users.noreply.github.com>
Co-authored-by: sunnycase <sunnycase@live.cn>
2024-03-05 12:01:46 +08:00
bdaf0b12c6
Fix in_ci env judge with bool type. ( #1171 )
...
* Fix in_ci env judge with bool type.
* Add fix for accuracy_test.
2024-02-22 20:21:20 +08:00
24fcfffac0
Fix the missing new line issue of dynamic shape. ( #1170 )
2024-02-22 10:31:41 +08:00
699f1980dc
Fix Melgan ( #1168 )
...
* update
* Apply code-format changes
---------
Co-authored-by: FusionBolt <FusionBolt@users.noreply.github.com>
2024-02-21 10:12:08 +08:00
a56268217e
Supprot npy for pytest ( #1164 )
...
* update
* Apply code-format changes
---------
Co-authored-by: FusionBolt <FusionBolt@users.noreply.github.com>
2024-02-04 10:48:52 +08:00
e7d0aa1d8e
Fix/dump_manager ( #1163 )
...
* fix dump manager
* Apply code-format changes
* update
* Apply code-format changes
* update
---------
Co-authored-by: curioyang <curioyang@users.noreply.github.com>
Co-authored-by: FusionBolt <RealFusionBolt@gmail.com>
Co-authored-by: FusionBolt <FusionBolt@users.noreply.github.com>
2024-02-01 18:01:04 +08:00
546d774700
Support multiple kmodels for accuracy_test. ( #1162 )
2024-01-29 17:40:44 +08:00
de6c89daa8
Fix Bucket ( #1159 )
...
* update
* update
* update
* fix
* fix
* fix swish evaluator
* update
* fix
* update
* update
* Apply code-format changes
* update
* fix transpose
* Add test cases for onnx Transpose.
* Apply code-format changes
---------
Co-authored-by: FusionBolt <FusionBolt@users.noreply.github.com>
Co-authored-by: zhangyang2057 <zhangyang@canaan-creative.com>
Co-authored-by: zhangyang2057 <zhangyang2057@users.noreply.github.com>
2024-01-24 14:03:25 +08:00
96477560f3
GNNE-1714:Fix/some opt ( #1158 )
...
* add Macro config method for processers
* update FoldNopBinary to support tensor
* restrict the Scenes to no broadcast
* add situation if lhs is var
* to avoid checkedShape error
---------
Co-authored-by: guodongliang <guodongliang@canaan-creative.com>
2024-01-18 18:00:39 +08:00
3331d2d84d
Feature/gem5 ( #1157 )
...
* layernrom and conv2d support float16
* update pip in ci
2024-01-18 09:58:42 +08:00
aa69f576ce
Fix Not Equal ( #1156 )
2024-01-09 12:08:23 +08:00
ebac64a5cf
GNNE-1714:Fix/search time egraph ( #1155 )
...
* change name method for getitem for lstm
* change search time to 120s
---------
Co-authored-by: guodongliang <guodongliang@canaan-creative.com>
2024-01-05 18:47:25 +08:00
52fe13b0e3
add clear device mem pool util ( #1153 )
...
* add clear device mem pool util
* Apply code-format changes
* rename api
* Apply code-format changes
---------
Co-authored-by: curioyang <curioyang@users.noreply.github.com>
2023-12-27 14:10:51 +08:00
907ae20755
Studio Doc ( #1152 )
...
* update size
* add doc
2023-12-26 13:52:02 +08:00
4f060909bb
Fix Studio ( #1151 )
...
* fix load calib
* update
* update
* Apply code-format changes
* update
---------
Co-authored-by: FusionBolt <FusionBolt@users.noreply.github.com>
2023-12-25 18:22:59 +08:00
a0e3ff5c90
adjust for processor counter ( #1150 )
...
Co-authored-by: guodongliang <guodongliang@canaan-creative.com>
2023-12-25 11:52:12 +08:00
c232f96948
GNNE-1980 Xpu Sram Codegen ( #1148 )
...
* fix
* add TryGetNonUniformDividedSlice
* add floordiv and ceildiv
* fix bug
* refactor unfold block
* add method
* fix nest builder
* TensorConst support DistributedType
* support distrubute type
* fix typeinfer
* Apply code-format changes
* fix build
---------
Co-authored-by: huochenghai <huochenghai@canaan-creative.com>
Co-authored-by: zhen8838 <zhen8838@users.noreply.github.com>
2023-12-22 17:06:21 +08:00
d355a380f7
Fix Reshape ( #1147 )
2023-12-22 13:56:41 +08:00
f7b4c7c18e
Refactor Quantize ( #1144 )
...
* Fix memory leak: Avoid disposing python stream in finalizer thread
* refactor
* update
* update
* update
* speedup but a little leak
* update
* fix
* clear print
* update
* update
* update
* update
* replace tensor with span
* remove print
* Apply code-format changes
* update
* Apply code-format changes
* update
---------
Co-authored-by: sunnycase <sunnycase@live.cn>
Co-authored-by: FusionBolt <FusionBolt@users.noreply.github.com>
2023-12-20 13:55:12 +08:00
3a533466ee
Feature/k800 qemu ( #1145 )
...
* add cmake of k800 toolchain
* add xpu in config.toml
* distributed type for where
* fix noptq
2023-12-20 10:17:56 +08:00
980f97747c
Fix memory leak: Avoid disposing python stream in finalizer thread ( #1142 )
2023-12-13 15:48:01 +08:00
c4f7b5fb14
Fix ordering of PytestCalibrationDatasetProvider ( #1143 )
2023-12-13 14:09:15 +08:00
67b77b3b9d
Add accuracy test. ( #1140 )
...
* Add accuracy test.
* Move if_quant/w_quaint into remark.
* enlarge qsize for CI server and reset postprocess result for nncase.
* Add '\n' for each item in remark.
2023-12-08 17:34:47 +08:00
778d723a5d
Fix/retinaface similarity bugs ( #1141 )
...
* change cost for resize image
* add strict infor for combineQuantize
* refuse to concat quant if no quant
* add tuple is current decison
---------
Co-authored-by: guodongliang <guodongliang@canaan-creative.com>
2023-12-04 16:28:45 +08:00
af513996e0
Fix Catch for Compile ( #1137 )
...
* fix bug
* fix
* update
* update
* resolve review
* update
* fix try catch
2023-11-29 18:00:58 +08:00
2b686b4484
Fix Studio ( #1136 )
...
* fix bug
* fix
* update
* update
* resolve review
* update
2023-11-29 15:57:58 +08:00
87aa40794d
add rvv_memcpy && opt profile ( #1135 )
...
* add rvv_memcpy && opt profile
* Apply code-format changes
* fix
* fix kernel test
* rename
---------
Co-authored-by: yanghaoqi <yanghaoqi_intern@canaan-creative.com>
Co-authored-by: curioyang <curioyang@users.noreply.github.com>
2023-11-28 16:28:07 +08:00
b3d8ae9beb
Nncase studio ( #1125 )
...
* update
* fix csproj
* tmp commit
* refactor finish
* fix build
* update
* update
* update
* add ShapeBucketView
* update
* fix cancel and simulate time
* update font family and progress bar
* fix build
* fix
* Apply code-format changes
* move file location
* fix
* update page switch logic
* resolve feedback
* close DBus
* update
* add some comment
* update
* fix build
* update
* update
* Apply code-format changes
* add CustomValidation
* add validate
* add some validate
* Apply code-format changes
* resolve review
* fix msg box
* add CompileConfig
* fix config export and import
* Apply code-format changes
* update preprocess
* Apply code-format changes
* add publish and upload for nncase-studio
* change artifacts name
* fix
---------
Co-authored-by: FusionBolt <FusionBolt@users.noreply.github.com>
2023-11-24 14:34:12 +08:00
bfa9c4da98
add onnx-op-support-md file ( #1081 )
...
* fix
* fix
* fix
* fix
* fix
* add kernel test codecov report
* fix
* fix
* fix
* update onnx_ops.md
* update onnx_ops.md
* fix tensor_ops
* fix
* fix
* fix
* update onnx_ops.md
* update onnx_ops.md
* update onnx_ops.md
* fix pytest error
* fix
* small fix
---------
Co-authored-by: hejunchao <hejunchao@canaan-creative.com>
Co-authored-by: Curio Yang <curioyhq@gmail.com>
Co-authored-by: sunnycase <sunnycase@live.cn>
2023-11-23 10:42:48 +08:00
c326fe3d87
add CheckedShape for DistributedType ( #1133 )
...
* add CheckedShape for DistributedType
* mod by 1 is not nop binary
* fix bug about export json in kld
* change unit test mode to NoClip
* fix warning
---------
Co-authored-by: guodongliang <guodongliang@canaan-creative.com>
2023-11-22 17:22:38 +08:00
352b41d091
fix metadata ( #1132 )
2023-11-21 17:28:25 +08:00
f6e4674ca2
fix LN evaluator ( #1131 )
2023-11-21 16:51:38 +08:00
2571d16baf
add missing meta information for quant ( #1130 )
...
Co-authored-by: guodongliang <guodongliang@canaan-creative.com>
2023-11-20 18:22:44 +08:00
3820af1eaf
Fix/melgan ( #1126 )
...
* add FoldDilatedConv2D
* add SwapBinaryArg
* update merge strategy(static shape, merge multiuser for normal, rebuild split from bucket, extract bucket condition)
* fix SpaceToBatch
* fix ShapeBucket
* fix
* add test
* fix
* fix
* fix
* fix marker
* fix some bug
* fix some bug
* add FoldTransposeActTranspose
* update
* format and fix some build
* fix build
* Apply code-format changes
* remove unused
* fix test
* Apply code-format changes
* fix
* fix test
* fix importer
* fix something
* update
* update
---------
Co-authored-by: FusionBolt <FusionBolt@users.noreply.github.com>
2023-11-17 14:47:58 +08:00
5364b0153e
Fix runtime ( #1115 )
...
* fix runtime
* Apply code-format changes
---------
Co-authored-by: FusionBolt <FusionBolt@users.noreply.github.com>
2023-11-16 09:48:43 +08:00
050de8ed39
GNNE-1974: Feature/non uniform ( #1128 )
...
* add split
* fix tests
2023-11-16 08:47:17 +08:00
734cce01c6
GNNE-1714:add norm op name ( #1127 )
...
* add norm op name
* opt realization
---------
Co-authored-by: guodongliang <guodongliang@canaan-creative.com>
Co-authored-by: FusionBolt <59008347+FusionBolt@users.noreply.github.com>
2023-11-15 17:54:26 +08:00
e4319655e5
Feature/remove apply ( #1124 )
...
* remove apply
* Apply code-format changes
* support onnx op 13
* Apply code-format changes
* fix bug
---------
Co-authored-by: yanghaoqi <yanghaoqi_intern@canaan-creative.com>
Co-authored-by: curioyang <curioyang@users.noreply.github.com>
2023-11-15 15:40:29 +08:00
e39a9d23a8
GNNE-1714:Fix/fix transform about quant ( #1123 )
...
* support dim5 slice
* add metadata for some transform
---------
Co-authored-by: guodongliang <guodongliang@canaan-creative.com>
2023-11-13 15:13:03 +08:00
381707215f
shutdown FoldPrePostReshapeSoftmax transform ( #1121 )
...
Co-authored-by: guodongliang <guodongliang@canaan-creative.com>
2023-11-09 09:55:53 +08:00
Curio Yang
huochenghai
zhangyang2057
sunnycase
FusionBolt
uranus0515
郑启航
Curio Yang
HeJunchao