박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Tue, 4 Dec 2018 01:36:51 +0000 (10:36 +0900)]
[tfldump] Empty tflite Model dump (#2466)
This will introduce empty tflite::Model dump as a place holder
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
[nnc] MeanSoftBackend (#2367)
This adds Support for reduction with Mean to C++ soft backend.
The implementation is a reference one from TFLite.
Signed-off-by: Andrei Shedko <a.shedko@partner.samsung.com>
[nnc] Support non-unit batch in deconv (#2471)
Added support of Batch dimension in Deconv operation
* shape inference
* interpreter
* c++ softBackend
Signed-off-by: Andrei Shedko <a.shedko@partner.samsung.com>
[nnc] Fix pool operation round mode (#2470)
* Add round mode to Pool operations for each framework
* Fix output shape for pooling operation
* Fix strides on Pool op shape inference
Signed-off-by: Pavel Iliutchenko <p.iliutchenk@samsung.com>
[nnc] Implementation of Transpose operation (#2464)
* Add `Transpose` operation to ModelIR;
* Support `Transpose` operation in interpreter and soft backend.
Signed-off-by: Sergei Barannikov <s.barannikov@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Mon, 3 Dec 2018 07:25:21 +0000 (16:25 +0900)]
[tfldump] Load tflite file (#2465)
This will add tflite loader part of tfldump
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Mon, 3 Dec 2018 02:20:42 +0000 (11:20 +0900)]
[tfldump] Add tflite schema flatbuffers (#2445)
* [tfldump] Add tflite schema flatbuffers
This will add tflite schema and flatbuffers converter to generate the header
Empty Model.h is added to check the build
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
* cmake changes
* rename files
* another rename
[nnc] Set operations names based on layer names in Caffe importer (#2463)
Refactor Caffe importer to set Operation names based on layer names (previously names were set based on output blob names).
Signed-off-by: Sergei Barannikov <s.barannikov@samsung.com>
[nnc] Support batch axis in acl backend (#2462)
- Support batch axis
- Add more checks on supported data formats
- Add assert in transposeShape and related fixes
Signed-off-by: Efimov Alexander <a.efimov@samsung.com>
[nnc] Utility to generate input tensors (#2143)
Implemented utility that generates input data in the hdf5 format.
Signed-off-by: Timur Ablyazimov <t.ablyazimov@samsung.com>
[nnc][utils] TFLIte model generator. Add operations. (#2383)
Add support of generating nodes with several random
inputs which are consistent by tensor's shape.
Add operations:
- DepthwiseConv2D
- MaxPool/AvgPool
- Softmax
- fullyConnected
Signed-off-by: i-kulagin <i.kulagin@samsung.com>
[nnc] Remove redundant CLCopy layer from inference sequence (#2449)
Remove buggy CLCopy layer from inference, since it is not needed, and corrupts data
Signed-off-by: Efimov Alexander <a.efimov@samsung.com>
[nnc] Support asimmetrical paddings in tflite models in ACL backend (#2447)
Inference asimmetrical paddings for "Same" padding type on acl backend
Signed-off-by: Efimov Alexander <a.efimov@samsung.com>
[nnc] Fix special case of bias layer after fc layer in ACL backend (#2448)
Handle shape of bias weights with respect to previous operation.
Need to use different shape for operations after fully connected layer,
because it restores batch dimension in output tensor
Signed-off-by: Efimov Alexander <a.efimov@samsung.com>
[nnc] Fix importer test (#2457)
Fixed exit code in importer test, added a version check so that the test
is not run on older TF
Signed-off-by: Andrei Shedko <a.shedko@partner.samsung.com>
Павел Ильютченко/AI Tools Lab /SRR/Engineer/삼성전자 [Fri, 30 Nov 2018 10:25:13 +0000 (13:25 +0300)]
[nnc] Support Pad operation on soft backend (#2364)
* Add serialization on softbackend
* Add deserialization on softbackend
* Add cpp_pad snippet
Signed-off-by: Pavel Iliutchenko <p.iliutchenk@samsung.com>
박종현/동작제어Lab(SR)/Staff Engineer/삼성전자 [Fri, 30 Nov 2018 08:19:59 +0000 (17:19 +0900)]
[enco/tfl] Validate stride (#2455)
This commit introduces basic tfl model validation infrastructure to tfl
frontend.
The current implementation supports 'stride' validation for the following
operations:
- Conv2D
- DepthwiseConv2D
- MaxPool2D
- AveragePool2D
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
박천교/동작제어Lab(SR)/Engineer/삼성전자 [Fri, 30 Nov 2018 04:35:08 +0000 (13:35 +0900)]
[encodump] Fix a bug not generating comma (#2421)
Now generating comma separated list output for size more than 1.
Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Fri, 30 Nov 2018 02:11:15 +0000 (11:11 +0900)]
[tflchef] Store op codes only once (#2401)
* [tflchef] Store op codes only once
This will change tflchef to store operator codes only once in the table
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
* apply comments
[nnc] Perform shape inference at construction time (#2399)
* Move shape inference functionality from separate class to constructors of individual operations;
* First dimension of input is no longer removed by the importers;
* Adjust ModelIR, soft backend and interpreter to correctly work with non-stripped first dimension;
* Minor coding style fixes and comments.
Signed-off-by: Sergei Barannikov <s.barannikov@samsung.com>
[nnc] Added reducemean to modelir (#2234)
- Added ReduceMean to ModelIR (Mean in tflite, not present in caffe)
The functionality allows one to support other reduction operations
rather effortlessly (in interpreter).
- Added tests for shape Inference.
Signed-off-by: Andrei Shedko <a.shedko@partner.samsung.com>
[nnc] ConstantOp, shape inference, model analyze, serialize, first artifact initial versions were implemented. (#2394)
New ConstantOp was added: at the moment it's used for input constant tensors like weights, etc.
Shape inference, model analyzing, etc, are working enough to be able to create the first artifact.
Signed-off-by: Andrew V. Tischenko <a.tischenko@partner.samsung.com>
[nnc] Make Mobile and Inception-v3 working (#2416)
These commit starts a sequence of commits introducing the debugged version of the ACL backend operations used in the Mobilenet and Inception-v3 networks.
Signed-off-by: Timur Ablyazimov <t.ablyazimov@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Thu, 29 Nov 2018 06:12:59 +0000 (15:12 +0900)]
[tfldump] Tool to dump TFlite file (#2444)
This will introduce empty project for tfldump tool to dump TFlite file to console in text form.
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Thu, 29 Nov 2018 05:50:41 +0000 (14:50 +0900)]
[enco/tfl/frontend] Creation of context ob outside of loop (#2442)
This commit moves creation of context ob outside of loop.
(putting creation code inside loop creates same object over and over)
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
박종현/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 29 Nov 2018 05:25:48 +0000 (14:25 +0900)]
[enco] Support constant folding (#2441)
This commit introduces constant folding pass into the enco compilation
pipeline.
Note that the current implementation supports folding over "copy"
instruction only.
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
박종현/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 29 Nov 2018 01:30:05 +0000 (10:30 +0900)]
[enco] Properly insert setOperandValue calls (#2439)
With the current implementation, it is possible that some setOperandValue
calls are missing during ANN IR construction.
With this commit, enco now inserts setOperandValue call for every operand
with constant weight.
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
[nnc] Add Resize Nearest Neighbor (#2315)
Added Resize Nearest Neighbor to tflite importer and interpreter.
Added shape inference tests some cases.
The op allows adding more resize types by just adding new values to the enum and backends.
The corresponding Op in ONNX is Upsample and it can be supported without modifications
Signed-off-by: Andrei Shedko <a.shedko@partner.samsung.com>
Add Test Round 2.1 in System Test Document (#2329)
* added tests for Tizen and SmartMachine OS
Signed-off-by: Roman Rusyaev <r.rusyaev@samsung.com>
Vladimir Plazun/AI Tools Lab /SRR/Engineer/삼성전자 [Wed, 28 Nov 2018 14:14:24 +0000 (17:14 +0300)]
[nnc] Remove redundant interpreter options (#2425)
- Remove cli options for input/output node, which are used only in interpreter
- fix code style in interpreter_pass.cpp
Signed-off-by: Vladimir Plazun <v.plazun@partner.samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 28 Nov 2018 04:48:28 +0000 (13:48 +0900)]
[enco/tfl/frontend] change names : `conv..` to `dconv...` (#2431)
By refactoring, names with `conv...` was changed to `dconv...`.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 28 Nov 2018 04:25:06 +0000 (13:25 +0900)]
[enco/tfl/frontend] test for DepthwiseConv2D for mobilenet (#2428)
Test case for IFM width = 2 * OFM width, IFM height = 2 * OFM height, SAME padding, stride 2
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 28 Nov 2018 03:45:35 +0000 (12:45 +0900)]
[enco/tfl/frontend] DepthwiseConv2D (same padding, ifm h=ofm h, ifm w=ofm w) (#2391)
* [enco/tfl/frontend] DepthwiseConv2D
DepthwiseConv2D for enco tflite frontend. Tested with stride=[1,1]
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
* pr fix: enter
* pr fix: removed swap, renaming some value, added asserts
* pr fix: Replaced "comment" with "#"
* pr fix: Fix unefficient declaration of new_shape
박종현/동작제어Lab(SR)/Staff Engineer/삼성전자 [Wed, 28 Nov 2018 03:43:49 +0000 (12:43 +0900)]
[enco] Delegate Concat to NN runtime (#2426)
With this commit, enco backend delegates ConcatF to NN runtime
instead of lowering it as a sequence of Shuffle.
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 28 Nov 2018 00:19:00 +0000 (09:19 +0900)]
[tflchef] Replace test files with cmake file GLOB (#2380)
This will replace individual test list preparation with file GLOB
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Tue, 27 Nov 2018 06:00:18 +0000 (15:00 +0900)]
[tflchef/rev] support quantization (#2402)
* [tflchef/rev] support quantization
This will enable reading quantization from tflite and store to tflchef
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
* remove space
* add note
* fix typo
* remove invalid '> 0'
남궁석/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 27 Nov 2018 04:07:22 +0000 (13:07 +0900)]
[tflchef] Support quantization parameters (#2379)
* [tflchef] Support quantization parameters
This commit will introduce quantization parameters and add related tests
Signed-off-by: Seok NamKoong <sk.namkoong@samsung.com>
* modify convention and data
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Tue, 27 Nov 2018 04:05:44 +0000 (13:05 +0900)]
[tflchef/rev] Write reshape explicit values (#2418)
This will change tflchef-reverse to provide explicit values of shape information for the operand used by reshape operator
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Tue, 27 Nov 2018 02:40:08 +0000 (11:40 +0900)]
[enco/tfl/frontend] Padding when top == bottom -1 or left = right-1 (#2411)
* [enco/tfl/frontend] Padding when top == bottom -1 or left = right-1
get_padding() now calculates padding when top == bottom -1 or left = right-1
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
* pr fix: calculate -> compute, ifm -> in_size
박종현/동작제어Lab(SR)/Staff Engineer/삼성전자 [Tue, 27 Nov 2018 02:35:44 +0000 (11:35 +0900)]
[enco] Do NOT reduce identical objects with side effect (#2419)
* [enco] Do NOT reduce identical objects with side effect
Update on one object may have a side effect through its backing bag. So,
reduction in identical objects with side effect results in incorrect
compilation artifact.
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
* Update the examples in the comment
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Mon, 26 Nov 2018 23:51:06 +0000 (08:51 +0900)]
[tflchef] Remove "comment" attribute (#2412)
"comment" is removed since "#" is comment in protobuffer schema.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
Ivan Vagin/AI Tools Lab /SRR/Engineer/삼성전자 [Mon, 26 Nov 2018 23:46:11 +0000 (02:46 +0300)]
Download pytorch sources on cmake stage (#2288)
Download pytorch sources on cmake stage
Signed-off-by: Ivan Vagin <ivan.vagin@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Mon, 26 Nov 2018 09:53:22 +0000 (18:53 +0900)]
[enco] Replaced "comment" to "#" in tfl test recipes (#2403)
By this commit, "comment" field in tfl test recipes was replaced to "#"
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
박종현/동작제어Lab(SR)/Staff Engineer/삼성전자 [Mon, 26 Nov 2018 09:46:24 +0000 (18:46 +0900)]
[tflchef] Support explicit initialization over INT32 tensors (#2408)
* [tflchef] Support explicit initialization over INT32 tensors
This commit introduces 'explicit' filler which allows users to
explicitly specifies the values of operands.
Currently, 'explicit' filler is available for INT32 types.
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
* Use -1 as a new_shape parameter
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Mon, 26 Nov 2018 02:10:32 +0000 (11:10 +0900)]
[encodump] Prints address of coco::op or coco::instr (#2396)
Previously, running encodump to print Copy prints addr of coco::Producer. Now prints coco::Instr. This applied to many ops and insts.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Mon, 26 Nov 2018 01:36:15 +0000 (10:36 +0900)]
[encodump] Added ReLU6 (#2395)
Withi this commit, encodump prints information about Relu6.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
Vladimir Plazun/AI Tools Lab /SRR/Engineer/삼성전자 [Fri, 23 Nov 2018 18:36:58 +0000 (21:36 +0300)]
[nnc] Implement elementwise div operation (#2363)
Added support of div elementwise:
- in tflite importer
- in interpreter
- in soft backend using Eigen
Signed-off-by: Vladimir Plazun <v.plazun@partner.samsung.com>
Vladimir Plazun/AI Tools Lab /SRR/Engineer/삼성전자 [Fri, 23 Nov 2018 17:53:52 +0000 (20:53 +0300)]
[nnc] Add support for freestanding activation functions in tflite importer (#2385)
While tflite have fused activation functions, some networks use activation functions as separate operations
Signed-off-by: Vladimir Plazun <v.plazun@partner.samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Fri, 23 Nov 2018 08:01:07 +0000 (17:01 +0900)]
[tflchef/reverse] Enable INT32 type (#2393)
This will enable tflchef-reverse to handle INT32 type
- Needed to handle Reshape operator from MobileNet
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
이상규/동작제어Lab(SR)/Principal Engineer/삼성전자 [Fri, 23 Nov 2018 04:07:20 +0000 (13:07 +0900)]
[tflchef] Update depthwiseconv2d recipe (use relu6 activation) (#2322)
depthwiseconv2d recipe will use relu6 activation.
- Mobilenet uses relu6 activation.
- It could be used as a test for relu6 activation.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Fri, 23 Nov 2018 04:06:01 +0000 (13:06 +0900)]
[tflchef] Support int32 tensortype (#2388)
This will fix to support int32 tensortype with introducing as_tflite_tensortype()
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Fri, 23 Nov 2018 01:37:19 +0000 (10:37 +0900)]
[tflchef] enable ReLU6 conversion (#2381)
This will enable ReLU6 in as_tflchef_activation()
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Fri, 23 Nov 2018 01:36:42 +0000 (10:36 +0900)]
[enco/frontend] Empty DepthwiseConv2D (#2387)
This will add placeholder for DepthwiseConv2D in frontend
- to check loading mobilenet slim model
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Fri, 23 Nov 2018 01:04:38 +0000 (10:04 +0900)]
[enco/tfl/frontend] ReLU6 Activation Layer (#2359)
* [enco/tfl/frontend] ReLU6 Activation Layer
This commit is for ReLU6 Activation Layer.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
* Pr fix: add code to op/Activation.cpp
[nnc] Support Pad operation on TFLite Importer and Interpreter (#2330)
* Support Pad op in TFLite imported and interpreter
Signed-off-by: Pavel Iliutchenko <p.iliutchenk@partner.samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Thu, 22 Nov 2018 08:20:04 +0000 (17:20 +0900)]
[tflchef] Add DepthwiseConv2D for reverse (#2373)
This will add DepthwiseConv2D Op for tflchef-reverse
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Thu, 22 Nov 2018 07:59:01 +0000 (16:59 +0900)]
[tflchef] Add Conv2D for reverse (#2375)
* [tflchef] Add Conv2D for reverse
This will add Conv2D Op for tflchef-reverse
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
* add comment
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Thu, 22 Nov 2018 07:48:01 +0000 (16:48 +0900)]
[tflchef] Add MaxPool2D for reverse (#2376)
This will add MaxPool2D Op for tflchef-reverse
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Thu, 22 Nov 2018 07:46:43 +0000 (16:46 +0900)]
[tflchef] Add ReLU6 for reverse (#2374)
This will add ReLU6 Op for tflchef-reverse
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Thu, 22 Nov 2018 05:58:55 +0000 (14:58 +0900)]
[tflchef] Add ReLU for reverse (#2377)
This will add ReLU Op for tflchef-reverse
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Thu, 22 Nov 2018 05:19:21 +0000 (14:19 +0900)]
[tflchef] Add Concat for reverse (#2370)
This will add Concatenation Op for tflchef-reverse
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Thu, 22 Nov 2018 05:12:23 +0000 (14:12 +0900)]
[tflchef] Add Reshape for reverse (#2371)
This will add Reshape Op for tflchef-reverse
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Thu, 22 Nov 2018 05:08:03 +0000 (14:08 +0900)]
[enco] Splitting Frontend.cpp into Op/*.cpp and more (#2362)
* [enco] Splitting Frontend.cpp into Op/*.cpp and more
Through this commit, Frontend.cpp will be refactored into several files.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
* pr fix: remove useless "tflite::"
* pr fix: Split h file into h and cpp file
* pr fix: remove wrong test files
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Thu, 22 Nov 2018 01:33:51 +0000 (10:33 +0900)]
[tflchef] Fix FlatBufferIntArrayToVector (#2369)
* [tflchef] Fix FlatBufferIntArrayToVector
This will rename FlatBufferIntArrayToVector to as_index_vector like enco/frontend and correct type of caller
- incorrect type may cause a problem in reading minus values in shape info
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
* fix to typename
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Thu, 22 Nov 2018 01:33:22 +0000 (10:33 +0900)]
[tflchef] Prepare adding Ops with AveragePool2D (#2368)
This will add AveragePool2D operator handler and needed changes to test.
Also place holders for other operators to prevent merge conflict.
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 21 Nov 2018 22:18:40 +0000 (07:18 +0900)]
[tflchef] generate_recipe for each operators (#2361)
This will fill handlers for each operators
- forward filler settings for the operator
- tflchef operator build and I/O settings for corresponding tflite operator
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
[nnc] Fix deconv unittest build (#2366)
Fix deconv unit test build and adjust float epsilon
Signed-off-by: Efimov Alexander <a.efimov@samsung.com>
[nnc][utils][model_generator] Add converter from tree to tflite. (#2230)
Implemented converter from randomly generated Tree representation to TFLite model.
Signed-off-by: i-kulagin i.kulagin@samsung.com
[nnc] Initial support of various operations in ONNX frontend (#2286)
Operations supported in ONNX importer:
- Conv2d
- BatchNorm
- Pool
- scale
- dropout
- Elementwise operations (add, sum, mul, max)
- Gemm
Signed-off-by: Andrew V. Tischenko <a.tischenko@partner.samsung.com>
[nnc] Conv transpose in SoftBackend (#2290)
Implements Deconvolution in Soft Backend, caffe style transfer model is fully supported.
Signed-off-by: Andrei Shedko a.shedko@partner.samsung.com
[nnc] Add `shape` parameter to constructors of operations that cannot infer shapes of their outputs (#2338)
* Add `shape` parameter to constructor of `VariableOp`;
* Add `shape` parameter to constructor of `ReshapeOp`.
Signed-off-by: Sergei Barannikov <s.barannikov@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 21 Nov 2018 10:31:57 +0000 (19:31 +0900)]
[tflchef] Implement generate_recipe (#2360)
* [tflchef] Implement generate_recipe
This implements partial of generate_recipe() for operators, operands and inputs,outputs.
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
* remove unused
* reorder a line
* fix typo
* need check quantization
* fix typo
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 21 Nov 2018 10:25:15 +0000 (19:25 +0900)]
[tflchef] Introduce TFliteOpChef and Registry (#2357)
This will introduce TFliteOpChef base class for Operator handlers for translation and
TFliteOpRegistry for registeration of handlers.
All headers are included in RecipeChef to check compilation and as it will be used later codes.
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 21 Nov 2018 09:51:18 +0000 (18:51 +0900)]
[tflchef] Adds "comment" into recipe (#2356)
Now, recipe file can have "comment" field, which is ignored by tflchef.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 21 Nov 2018 08:15:16 +0000 (17:15 +0900)]
[tflchef] Introduce TFliteImport (#2355)
* [tflchef] Introduce TFliteImport
This will introduce TFliteImport class that provides needed information from tflite::Model to build recipe
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
* add brief
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 21 Nov 2018 05:11:59 +0000 (14:11 +0900)]
[enco/tfl/frontend] Modified wrong test number (#2354)
Conv_003 exists without Conv_002. So Conv_003 is renamed to Conv_002.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 21 Nov 2018 05:11:09 +0000 (14:11 +0900)]
[tflchef] Introduce convert helpers (#2351)
This will introduce helper functions to convert tflite types to tflchef types
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 21 Nov 2018 05:10:39 +0000 (14:10 +0900)]
[tflchef] Store ModelRecipe to file (#2352)
This will implement storing ModelRecipe to a file
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 21 Nov 2018 05:10:19 +0000 (14:10 +0900)]
[enco/tfl/frontend] Bug fix on padding (#2353)
IFM is NHWC format but the padding code was written assuming that IFM is NCHW.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 21 Nov 2018 04:44:45 +0000 (13:44 +0900)]
[enco/tfl/frontend] Refactoring RELU activation (#2313)
* [enco/tfl/frontend] Refactoring RELU activation
RELU activation is commonly used by many operations. So this activation is extracted as a separate method.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
* returning ifm
* ofm as param
* remove empty line
* pr fix: remove output param
* pr fix : comment, renaming vars, etc.
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 21 Nov 2018 02:51:21 +0000 (11:51 +0900)]
[tflchef] Call tflchef from Driver (#2350)
This will make Driver call tflchef tflite load and save functions
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 21 Nov 2018 02:18:49 +0000 (11:18 +0900)]
[tflchef] RawModel loader (#2347)
This will introduce TFlite raw model loader
- copied from enco/frontend
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 21 Nov 2018 02:18:20 +0000 (11:18 +0900)]
[tflchef] Introduce entry points for tflchef-reverse (#2345)
* [tflchef] Introduce entry points for tflchef-reverse
This will introduce two empty functions for tflchef-reverse tool
- CMakeLists.txt file to build a library
- This library will read TF lite file and drop tflchef recipe file
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
* change header ifdef name
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 21 Nov 2018 01:19:12 +0000 (10:19 +0900)]
[tflchef] Introduce reverse tool driver (#2346)
[tflchef] Introduce reverse tool driver
[nnc] fix static analysis warning and add missed break (#2287)
- Add assert, which protects us from passing a null pointer to a function.
- Add break statement into switch in tflite_importer.
Signed-off-by: Pavel <p.fattakhov@partner.samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Tue, 20 Nov 2018 09:52:43 +0000 (18:52 +0900)]
[tflchef] Add Conv2D test recipe (#2340)
This will add Conv2D test recipe for tflchef.
- there was existing readme to show readme example works but there was no conv2d test
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
[nnc] Make constructors of `Operations` accept other `Operation`s as arguments (#2292)
* Make constructors of `Operations` accept `Operation`s they depend on.
* Remove `connectInput` method, which became redundant.
Signed-off-by: Sergei Barannikov <s.barannikov@samsung.com>
[nnc] split model runner script into 4 scripts (#2285)
split model runner script into 4 scripts
Signed-off-by: Pavel <p.fattakhov@partner.samsung.com>
박세희/동작제어Lab(SR)/Principal Engineer/삼성전자 [Mon, 19 Nov 2018 08:17:54 +0000 (17:17 +0900)]
[enco] Fix typo (#2334)
This will fix typo in OwnedFileDescriptor comment
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
남궁석/동작제어Lab(SR)/Engineer/삼성전자 [Fri, 16 Nov 2018 02:06:49 +0000 (11:06 +0900)]
[tflchef] Add ReLU6 in activation (#2312)
This commit will add ReLU6 in activation
Signed-off-by: Seok NamKoong <sk.namkoong@samsung.com>
남궁석/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 15 Nov 2018 23:09:30 +0000 (08:09 +0900)]
[enco] Remove unused variable (#2311)
This commit will remove unused varuable
Signed-off-by: Seok NamKoong <sk.namkoong@samsung.com>
박종현/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 15 Nov 2018 23:08:48 +0000 (08:08 +0900)]
[nnkit] Link backend_support library with dl (#2310)
Backend support library uses dlopen internally, but is not linked with
dl currently.
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
박종현/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 15 Nov 2018 23:07:52 +0000 (08:07 +0900)]
[enco] Set subnet output correctly on branching (#2309)
The current implementation specifies a bag as a subnet output if the subnet
accesses it and this bag is accessed outside of this subnet.
It turns out that this algorithm may specify a network input as a subnet output
as it does not consider this bag is updated by this subnet.
This commit fixes this bug and adds a related test.
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
박종현/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 15 Nov 2018 23:07:05 +0000 (08:07 +0900)]
[tflchef] Use safemain (#2305)
Now, tflchef and tflchef-file uses the main function that safemain
provides.
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
Ivan Vagin/AI Tools Lab /SRR/Engineer/삼성전자 [Thu, 15 Nov 2018 16:16:31 +0000 (19:16 +0300)]
[nnc] Implemented caffe2 importer interface (#2233)
* Implemented caffe2 importer interface
* Minor fixes in driver sanity checks
* Inited caffe2 model dumper
Signed-off-by: Ivan Vagin <ivan.vagin@samsung.com>
[nnc] Remove redundant code (#2314)
Remove redundant code from c++ snippet that prevents artifact build
Signed-off-by: Efimov Alexander <a.efimov@samsung.com>
[nnc] Replace layout of operations in cpu backend (#2258)
Order of operations in artifact is changed to RPO to optimize usage of memory caches
Signed-off-by: Efimov Alexander <a.efimov@samsung.com>
[nnc] Tfl importer eltwise (#2208)
Added elementwise operations, Tanh and Transposed Conv to tfLite importer
Signed-off-by: Andrei Shedko <a.shedko@partner.samsung.com>
박종현/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 15 Nov 2018 06:35:02 +0000 (15:35 +0900)]
[coco] Deprecate FeatureObject create with Shape (#2302)
This commit removes create method that takes a FeatureShape from
ObjectManager class and revises related tests.
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Thu, 15 Nov 2018 06:33:14 +0000 (15:33 +0900)]
[enco/frontend] Relu activation layer for Conv2D (#2306)
* [enco/frontend] Relu activation layer for Conv2D
This commit adds Relu activation layer for Conv2D and test.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
* Check NONE for actuvation