{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":249615315,"defaultBranch":"main","name":"nntrainer","ownerLogin":"nnstreamer","currentUserCanPush":false,"isFork":false,"isEmpty":false,"createdAt":"2020-03-24T04:47:36.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/60992508?v=4","public":true,"private":false,"isOrgOwned":true},"refInfo":{"name":"","listCacheKey":"v0:1721957379.0","currentOid":""},"activityList":{"items":[{"before":"e02535234d190fa76c6c0ba4c35ef4c9129a4692","after":"9eb4c85865cee29601a24ba4cea5e3bade61b936","ref":"refs/heads/main","pushedAt":"2024-09-11T22:47:18.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[ Tizen7.0 ] Include some headers in -dev header for neuralnet.h\n\n- In the previous PR (77e56f1), neuralnet.h was included in dev package.\n- However, some headers were missing used in nueralnet.h\n- This PR adds headers which have dependency with neuralnet.h\n- This PR is tested whether it supports ReinforcementLearning app on\nTizen7.0\n\nSelf evaluation:\n\nBuild test: [X]Passed [ ]Failed [ ]Skipped\nRun test: [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Eunju Yang <ej.yang@samsung.com>","shortMessageHtmlLink":"[ Tizen7.0 ] Include some headers in -dev header for neuralnet.h"}},{"before":"77e56f150048c60def3a0174c2499447c10835fe","after":"e02535234d190fa76c6c0ba4c35ef4c9129a4692","ref":"refs/heads/main","pushedAt":"2024-09-06T01:30:51.000Z","pushType":"pr_merge","commitsCount":2,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[Tizen7.0] Tizen7.0 Backporting\n\n- This commit adds some updates for Tizen7.0 backporting\n- Type mismatch bug is fixed.\n- Unused variable is removed.\n- Missing header files are added in spec file.\n- spec file is updated\n\nSelf evaluation:\n\nBuild test: [X]Passed [ ]Failed [ ]Skipped\nRun test: [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Eunju Yang <ej.yang@samsung.com>","shortMessageHtmlLink":"[Tizen7.0] Tizen7.0 Backporting"}},{"before":"6d6e924cecdb93d4b10cb41a2bbbb76de817dd7b","after":"b7da89e527ec763fc28cf554e578b6ee8df32d3f","ref":"refs/heads/release/tizen7.0","pushedAt":"2024-09-05T22:55:50.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[ Tizen7.0 ] Include neuralnet.h in -dev header\n\n- Update the code to include `neuralnet.h` in -dev header.\n- Some applications, e.g., ReinforcementLearning uses `forwarding` and\n`backwarding` directly. To support it, this commit adds the header into\ndev package.\n\n**Self-evaluation:**\n1. Build test: [X]Passed [ ]Failed [ ]Skipped\n2. Run test:   [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Eunju Yang <ej.yang@samsung.com>","shortMessageHtmlLink":"[ Tizen7.0 ] Include neuralnet.h in -dev header"}},{"before":"656a231787c426e021f0c58479d2ffaabf3558b2","after":"77e56f150048c60def3a0174c2499447c10835fe","ref":"refs/heads/main","pushedAt":"2024-09-05T22:41:37.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[ Tizen7.0 ] Include neuralnet.h in -dev header\n\n- Update the code to include `neuralnet.h` in -dev header.\n- Some applications, e.g., ReinforcementLearning uses `forwarding` and\n`backwarding` directly. To support it, this commit adds the header into\ndev package.\n\n**Self-evaluation:**\n1. Build test: [X]Passed [ ]Failed [ ]Skipped\n2. Run test:   [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Eunju Yang <ej.yang@samsung.com>","shortMessageHtmlLink":"[ Tizen7.0 ] Include neuralnet.h in -dev header"}},{"before":"6d6e924cecdb93d4b10cb41a2bbbb76de817dd7b","after":"656a231787c426e021f0c58479d2ffaabf3558b2","ref":"refs/heads/main","pushedAt":"2024-08-30T02:20:59.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[bugfix] fix coverity issues\n\nThis PR resolves coverity issues in the ShortTensor class.\nReplace max_abs() implementation with maxValue() since the maximum absolute value of unsigned int equals to the maximum value.\n\n**Self-evaluation:**\n1. Build test: [X]Passed [ ]Failed [ ]Skipped\n2. Run test:   [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Donghyeon Jeong <dhyeon.jeong@samsung.com>","shortMessageHtmlLink":"[bugfix] fix coverity issues"}},{"before":"a2c3a29b85b0d3b083addb7815a4f401fc6127c0","after":"6d6e924cecdb93d4b10cb41a2bbbb76de817dd7b","ref":"refs/heads/release/tizen7.0","pushedAt":"2024-08-29T11:05:06.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[Tizen7.0] Tizen7.0 Backporting\n\n- This commit adds some updates for Tizen7.0 backporting\n- Type mismatch bug is fixed.\n- Unused variable is removed.\n- Missing header files are added in spec file.\n- spec file is updated\n\nSelf evaluation:\n\nBuild test: [X]Passed [ ]Failed [ ]Skipped\nRun test: [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Eunju Yang <ej.yang@samsung.com>","shortMessageHtmlLink":"[Tizen7.0] Tizen7.0 Backporting"}},{"before":"0c9c6d7e7c4124ea4fe5c498c259de3ae9b3fc08","after":"6d6e924cecdb93d4b10cb41a2bbbb76de817dd7b","ref":"refs/heads/main","pushedAt":"2024-08-29T10:33:54.000Z","pushType":"pr_merge","commitsCount":4,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[Tizen7.0] Tizen7.0 Backporting\n\n- This commit adds some updates for Tizen7.0 backporting\n- Type mismatch bug is fixed.\n- Unused variable is removed.\n- Missing header files are added in spec file.\n- spec file is updated\n\nSelf evaluation:\n\nBuild test: [X]Passed [ ]Failed [ ]Skipped\nRun test: [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Eunju Yang <ej.yang@samsung.com>","shortMessageHtmlLink":"[Tizen7.0] Tizen7.0 Backporting"}},{"before":"9ae7d2d7fc5002aa61ef8ecabd8a39a10aba7fea","after":"0c9c6d7e7c4124ea4fe5c498c259de3ae9b3fc08","ref":"refs/heads/main","pushedAt":"2024-08-29T04:08:20.000Z","pushType":"pr_merge","commitsCount":3,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[BUILD] Remove Flag and Add FLECIBLE PAGE Option\n\nRemove Flag on Android.mk\nAdd APP_SUPPORT_FLEXIBLE_PAGE_SIZE to Application.mk\n\n**Self evaluation:**\n1. Build test:\t [X]Passed [ ]Failed [ ]Skipped\n2. Run test:\t [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Donghak PARK <donghak.park@samsung.com>","shortMessageHtmlLink":"[BUILD] Remove Flag and Add FLECIBLE PAGE Option"}},{"before":"1377f1c55cbe9842f430b2851581e533c4be1303","after":"9ae7d2d7fc5002aa61ef8ecabd8a39a10aba7fea","ref":"refs/heads/main","pushedAt":"2024-08-29T04:08:04.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[ BUILD ] Add 16K shared lib package option for Android\n\nAndroid encourage to use 16KB package for the shared library. This PR\nadd the 16KB package option and also recommand to use ndk which is\nhigher or equal version of r27.\n\nResolves:\n\n**Self evaluation:**\n1. Build test:\t [X]Passed [ ]Failed [ ]Skipped\n2. Run test:\t [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: jijoong.moon <jijoong.moon@samsung.com>","shortMessageHtmlLink":"[ BUILD ] Add 16K shared lib package option for Android"}},{"before":"cc5ca33f9cbc6824c69796df1ddf3ecd8a0d563c","after":"1377f1c55cbe9842f430b2851581e533c4be1303","ref":"refs/heads/main","pushedAt":"2024-08-27T23:56:41.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[enhance] Using 64 bit for LayerKernel enum\n\nEnhanced LayerKernel enum and mask for 64-bit values\n\nSigned-off-by: Debadri Samaddar <s.debadri@samsung.com>","shortMessageHtmlLink":"[enhance] Using 64 bit for LayerKernel enum"}},{"before":"f373d8835a7dccd7f233b8e2f978b0540017fe50","after":"cc5ca33f9cbc6824c69796df1ddf3ecd8a0d563c","ref":"refs/heads/main","pushedAt":"2024-08-27T04:38:46.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"myungjoo","name":"MyungJoo Ham","path":"/myungjoo","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1109011?s=80&v=4"},"commit":{"message":"[Tensor] ShortTensor class with unsigned 16-bit integer\n\nIn this PR, a new type of tensor, the ShortTensor class, is designed explicitly for handling unsigned 16-bit integer data types.\nThis new tensor class aims to provide users with more options when working with various data types.\nNote that the ShortTensor class does not support mathematical operations like multiplication or addition.\n\n**Self-evaluation:**\n1. Build test: [X]Passed [ ]Failed [ ]Skipped\n2. Run test:   [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Donghyeon Jeong <dhyeon.jeong@samsung.com>","shortMessageHtmlLink":"[Tensor] ShortTensor class with unsigned 16-bit integer"}},{"before":"a8993a2b0a6a15628480eb91520d5c6aa6d65586","after":"f373d8835a7dccd7f233b8e2f978b0540017fe50","ref":"refs/heads/main","pushedAt":"2024-08-26T08:29:34.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"myungjoo","name":"MyungJoo Ham","path":"/myungjoo","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1109011?s=80&v=4"},"commit":{"message":"[ blas/bugfix ] Fix irrelevant function call\n\n- Since current function implementations are not using CBLAS params, should directly call function from cblas.h\n\n**Self evaluation:**\n1. Build test:\t [X]Passed [ ]Failed [ ]Skipped\n2. Run test:\t [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: skykongkong8 <ss.kong@samsung.com>","shortMessageHtmlLink":"[ blas/bugfix ] Fix irrelevant function call"}},{"before":"6623e3057bddbc03afea04c2ff2e02244db12289","after":"a8993a2b0a6a15628480eb91520d5c6aa6d65586","ref":"refs/heads/main","pushedAt":"2024-08-24T00:02:28.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"myungjoo","name":"MyungJoo Ham","path":"/myungjoo","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1109011?s=80&v=4"},"commit":{"message":"[bugfix] Resolve fp16 enabled build error\n\nThis PR resolves the build error after #2704 when enable_fp16 is true.\n\nThis fixes:\nblas_interface.cpp:141:9: error: ‘order’ was not declared in this scope\n  141 |   sgemv(order, TransA, M, N, alpha, A_, lda, X_, incX, beta, Y_, incY);\n      |         ^~~~~\n\n**Self-evaluation:**\n1. Build test: [X]Passed [ ]Failed [ ]Skipped\n2. Run test:   [ ]Passed [X]Failed [ ]Skipped\n\nSigned-off-by: Donghyeon Jeong <dhyeon.jeong@samsung.com>","shortMessageHtmlLink":"[bugfix] Resolve fp16 enabled build error"}},{"before":"0ceb7d9b5544abf8a78abae74f815f35e36c7fb0","after":"6623e3057bddbc03afea04c2ff2e02244db12289","ref":"refs/heads/main","pushedAt":"2024-08-22T22:46:53.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[ Tensor ] Remove CBLAS params from Tensor related files.\n\n- Remove cblas params from tensor related files since nntrainer is not fully-dependent on cblas anymore.\n- Letting tensors to be aware of Cblas related parameters is a nonsense at the first place.\n- CBLAS params will be declared only when functions from cblas is called.\n\n**Self evaluation:**\n1. Build test:     [X]Passed [ ]Failed [ ]Skipped\n2. Run test:     [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: skykongkong8 <ss.kong@samsung.com>","shortMessageHtmlLink":"[ Tensor ] Remove CBLAS params from Tensor related files."}},{"before":"75a2e932ae7cd036f2508a94068ca5ebe1891e7e","after":"0ceb7d9b5544abf8a78abae74f815f35e36c7fb0","ref":"refs/heads/main","pushedAt":"2024-08-22T22:43:29.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[ CAPI ] fix the Native API Ref Doc\n\nAdd MODULE in submodule name in doc file.\n\n**Self evaluation:**\n1. Build test:\t [X]Passed [ ]Failed [ ]Skipped\n2. Run test:\t [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: jijoong.moon <jijoong.moon@samsung.com>","shortMessageHtmlLink":"[ CAPI ] fix the Native API Ref Doc"}},{"before":"905a5b11decafa725928a0fcdd3a4e952a94ee0a","after":"75a2e932ae7cd036f2508a94068ca5ebe1891e7e","ref":"refs/heads/main","pushedAt":"2024-08-19T07:28:55.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[coverity] fix coverity issue\n\nThis PR resolves the coverity issues of resource leak, unreachable code, and missing break.\n\n**Changes proposed in this PR:**\n- use static arrays instead of dynamic allocation to avoid resource leaks.\n- remove unreachable code and add missing break statement.\n\n**Self-evaluation:**\n1. Build test: [X]Passed [ ]Failed [ ]Skipped\n2. Run test:   [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Donghyeon Jeong <dhyeon.jeong@samsung.com>","shortMessageHtmlLink":"[coverity] fix coverity issue"}},{"before":"1df526c15a36f75ea7d5b7a47589aaaadd89f728","after":"905a5b11decafa725928a0fcdd3a4e952a94ee0a","ref":"refs/heads/main","pushedAt":"2024-08-19T07:28:18.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[GPU/OPENCL] RMSNorm Accuracy Fix\n\nThe alpha values were not picked correctly.\n\nSigned-off-by: Thummala Pallavi <t.pallavi@samsung.com>","shortMessageHtmlLink":"[GPU/OPENCL] RMSNorm Accuracy Fix"}},{"before":"8e8196fd88bad7f56f072b00ff6d4b89dacd7687","after":"1df526c15a36f75ea7d5b7a47589aaaadd89f728","ref":"refs/heads/main","pushedAt":"2024-08-16T01:23:57.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[Layer] enhance ConcatLayer algorithms for efficient concatenation and split\n\nThis PR renovates reshape/concatenation algorithms to facilitate efficient concatenation and split in ConcatLayer.\n\nPreviously, dimension 2 (height) was set as a standard axis to operate concatenation.\nHowever, this causes an overhead by copying a tensor size of 1 when the concat dimension is 3 (width).\n\nThe new algorithm consolidates all dimensions to the first and last axes based on the concat dimension, sets the standard axis to be 3, and performs concat and split.\n\n**Changes proposed in this PR:**\n- Revise creating helper dimension logic in finalize().\n- Update forwarding() and calcDeriv() workflow to be efficient.\n- Add descriptions for the new concat algorithm.\n\nSigned-off-by: Donghyeon Jeong <dhyeon.jeong@samsung.com>","shortMessageHtmlLink":"[Layer] enhance ConcatLayer algorithms for efficient concatenation an…"}},{"before":"79a5179e9960a8b5adbcdf5255a2fb4893ed50c8","after":"8e8196fd88bad7f56f072b00ff6d4b89dacd7687","ref":"refs/heads/main","pushedAt":"2024-08-16T01:22:43.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[Tensor] Add optional output tensor for tensor concatenation\n\nThis PR adds an optional feature in Tensor::cat to pass the output tensor to the function.\nThis change allows the user-given tensor to store the result of the concatenation without creating a new tensor.\n\n**Changes proposed in this PR:**\n- Add optional argument output (the output tensor) to the cat function.\n- Add negative test cases for tensor concatenation.\n\n**Self-evaluation:**\n1. Build test: [X]Passed [ ]Failed [ ]Skipped\n2. Run test:   [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Donghyeon Jeong <dhyeon.jeong@samsung.com>","shortMessageHtmlLink":"[Tensor] Add optional output tensor for tensor concatenation"}},{"before":"3b11453bde0c6743c776c5f7f021e31ed074e82c","after":"79a5179e9960a8b5adbcdf5255a2fb4893ed50c8","ref":"refs/heads/main","pushedAt":"2024-08-16T01:20:17.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[Android] Support Android NDK r27 and higher\n\nThis PR enables NNTrainer to use Android NDK r27 to support compiling 16 KB-aligned shared libraries.\n\nWhile -fpu is ignored and -mfloat-abi option is not valid with AArch64 targets, removing these options has no effect on using current NEON instructions for armv8.2.\n\n**Self-evaluation:**\n1. Build test: [X]Passed [ ]Failed [ ]Skipped\n2. Run test:   [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Donghyeon Jeong <dhyeon.jeong@samsung.com>","shortMessageHtmlLink":"[Android] Support Android NDK r27 and higher"}},{"before":"0328879fede8554c88872dc9fd0c3c6d0e4fe0f9","after":"3b11453bde0c6743c776c5f7f021e31ed074e82c","ref":"refs/heads/main","pushedAt":"2024-08-12T01:23:22.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[Layer] Improve forwarding logic of ConcatLayer\n\nThis PR updates current ConcatLayer forwarding for faster computation.\n\n**Changes proposed in this PR:**\n- Utilize the Tensor::concat() operation to perform forwarding and replace manual mapping and copying.\n\n**Self-evaluation:**\n1. Build test: [X]Passed [ ]Failed [ ]Skipped\n2. Run test:   [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Donghyeon Jeong <dhyeon.jeong@samsung.com>","shortMessageHtmlLink":"[Layer] Improve forwarding logic of ConcatLayer"}},{"before":"aa1bddf04cf1f8fce653db4d391c1813444793b3","after":"0328879fede8554c88872dc9fd0c3c6d0e4fe0f9","ref":"refs/heads/main","pushedAt":"2024-08-09T04:48:56.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[Tensor] CharTensor class with signed 8-bit integer\n\nIn this PR, a new type of tensor, the CharTensor class, is designed explicitly for handling signed 8-bit integer data types that have already undergone quantization.\nThis new tensor class aims to provide users with more options when working with tensors and their respective data types.\nCurrently, the CharTensor class does not support mathematical operations like multiplication or addition. However, these features will be added in future updates.\n\n**Self-evaluation:**\n1. Build test: [X]Passed [ ]Failed [ ]Skipped\n2. Run test:   [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Donghyeon Jeong <dhyeon.jeong@samsung.com>","shortMessageHtmlLink":"[Tensor] CharTensor class with signed 8-bit integer"}},{"before":"99a2a3d3c0d3f9d78ebbfbe6ec0de9f06e9c3f7e","after":"aa1bddf04cf1f8fce653db4d391c1813444793b3","ref":"refs/heads/main","pushedAt":"2024-08-09T04:47:39.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[ matrix_transpose/bugfix ] Prevent reading/saving data from/to unallocated memory\n\n- Previous transpose kernel occasionally load/save unallocated memory, and then masked it.\n- Now, it does not read them at the first place, but load with for-loop\n- This would deteriorate speed of fp16 matrix transpose, but won't be dominant in total model latency\n\n**Self evaluation:**\n1. Build test:     [X]Passed [ ]Failed [ ]Skipped\n2. Run test:     [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: skykongkong8 <ss.kong@samsung.com>","shortMessageHtmlLink":"[ matrix_transpose/bugfix ] Prevent reading/saving data from/to unall…"}},{"before":"23c09837cb992d958a83cafdc7d6b7bff3d1eb15","after":"99a2a3d3c0d3f9d78ebbfbe6ec0de9f06e9c3f7e","ref":"refs/heads/main","pushedAt":"2024-08-09T04:46:40.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[ hgemm ] Generalize redundant micro hgemm kernel implementation\n\n- Previous implementation naively used fixed-sized ukernels for the K-direction accumulation.\n- Such kernels were excessively long, but had better performance than looping through single K-iteration.\n- However, recent test results have shown that justing stacking 4 K iters, and looping through such ukernel preserved the performance with better code readability.\n\n**Self evaluation:**\n1. Build test:     [X]Passed [ ]Failed [ ]Skipped\n2. Run test:     [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: skykongkong8 <ss.kong@samsung.com>","shortMessageHtmlLink":"[ hgemm ] Generalize redundant micro hgemm kernel implementation"}},{"before":"05f7b4a14a63fcf3e0b0b14367b49f1db48b2c2c","after":"23c09837cb992d958a83cafdc7d6b7bff3d1eb15","ref":"refs/heads/main","pushedAt":"2024-08-09T00:36:49.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[Layer] add Weight Layer\n\n- This layer contains only weights for building tensor-level graph\n\n**Self-evaluation:**\n1. Build test: [X]Passed [ ]Failed [ ]Skipped\n2. Run test:   [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Seungbaek Hong <sb92.hong@samsung.com>","shortMessageHtmlLink":"[Layer] add Weight Layer"}},{"before":"32d901cee3c913a27e95ce81923266f6de7a063b","after":"05f7b4a14a63fcf3e0b0b14367b49f1db48b2c2c","ref":"refs/heads/main","pushedAt":"2024-08-09T00:24:54.000Z","pushType":"pr_merge","commitsCount":4,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[ hgemm ] Apply hgemm util funcs at frequently used functions\n\n- get_prev_mltpl_of_2p_n is frequently used in many hgemm kernels.\n\n**Self evaluation:**\n1. Build test:     [X]Passed [ ]Failed [ ]Skipped\n2. Run test:     [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: skykongkong8 <ss.kong@samsung.com>","shortMessageHtmlLink":"[ hgemm ] Apply hgemm util funcs at frequently used functions"}},{"before":"8877d6abd390a35590dc3e20eb0e821e22ab686a","after":"32d901cee3c913a27e95ce81923266f6de7a063b","ref":"refs/heads/main","pushedAt":"2024-08-05T01:01:43.000Z","pushType":"pr_merge","commitsCount":2,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[refactor] Restructure getStringDataType function\n\nThis patch updates the getStringDataType function structure to utilize method overriding.\n\n**Self-evaluation:**\n1. Build test: [X]Passed [ ]Failed [ ]Skipped\n2. Run test:   [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Donghyeon Jeong <dhyeon.jeong@samsung.com>","shortMessageHtmlLink":"[refactor] Restructure getStringDataType function"}},{"before":"632c68ad451ebac98670d970e83aced6438ed31a","after":"8877d6abd390a35590dc3e20eb0e821e22ab686a","ref":"refs/heads/main","pushedAt":"2024-08-01T23:06:28.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"BUG FIX : Concat GPU Layer and CPU layer unittest cases name overlapping.\n\nModified the concat gpu testcases name in unittest_layers_concat_cl for differentiation with concat cpu testcases name.\n\n**Self evaluation:**\n1. Build test:   [X]Passed [ ]Failed [ ]Skipped\n2. Run test:     [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Niket Agarwal <niket.a@samsung.com>","shortMessageHtmlLink":"BUG FIX : Concat GPU Layer and CPU layer unittest cases name overlapp…"}},{"before":"9eed5b6b9d4654d10edd355b9419ba0975e84fb1","after":"632c68ad451ebac98670d970e83aced6438ed31a","ref":"refs/heads/main","pushedAt":"2024-08-01T01:52:20.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[Doc] NNTrainer Tool Utilization Guide\n\nThis PR adds a guide for executing unit tests on the Android device.\n\n**Self-evaluation:**\n1. Build test: [X]Passed [ ]Failed [ ]Skipped\n2. Run test:   [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Donghyeon Jeong <dhyeon.jeong@samsung.com>","shortMessageHtmlLink":"[Doc] NNTrainer Tool Utilization Guide"}},{"before":"c42102a60de6378c0e7ec46f16ffee09b41f72a0","after":"9eed5b6b9d4654d10edd355b9419ba0975e84fb1","ref":"refs/heads/main","pushedAt":"2024-08-01T01:51:45.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"jijoongmoon","name":"Jijoong Moon","path":"/jijoongmoon","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/23226860?s=80&v=4"},"commit":{"message":"[Android] Verify Android NDK Installation and Configuration\n\nThis patch checks if Android NDK is installed and configured before building using NDK in the Android test script.\n\n**Self-evaluation:**\n1. Build test: [X]Passed [ ]Failed [ ]Skipped\n2. Run test:   [X]Passed [ ]Failed [ ]Skipped\n\nSigned-off-by: Donghyeon Jeong <dhyeon.jeong@samsung.com>","shortMessageHtmlLink":"[Android] Verify Android NDK Installation and Configuration"}}],"hasNextPage":true,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"Y3Vyc29yOnYyOpK7MjAyNC0wOS0xMVQyMjo0NzoxOC4wMDAwMDBazwAAAASzvBey","startCursor":"Y3Vyc29yOnYyOpK7MjAyNC0wOS0xMVQyMjo0NzoxOC4wMDAwMDBazwAAAASzvBey","endCursor":"Y3Vyc29yOnYyOpK7MjAyNC0wOC0wMVQwMTo1MTo0NS4wMDAwMDBazwAAAASOoxiv"}},"title":"Activity · nnstreamer/nntrainer"}