v1.5.0

What's Changed

• Dev wukong by @wyfwyf-ux in #3
• feat: adjust stream and add Wukong model operators & tests by @wyfwyf-ux in #2
• Dev wukong _wide_deep by @wyfwyf-ux in #7
• feat: add musa leakyrelu/tanh ops and grad kernels with tests by @ssjcode in #12
• feat: add sign op by @greenhandF in #15
• Feat/pad by @ChaosAbyssF in #16
• Feat/round by @ChaosAbyssF in #17
• add less op by @greenhandF in #19
• Feat/random_standard_normal by @ChaosAbyssF in #13
• Mem manage by @tngchien in #24
• add bitwise and op by @greenhandF in #23
• Dev/variablev2 by @yaotianhang in #11
• add bitwise and op by @greenhandF in #26
• fix: fix expandim op by @tngchien in #27
• Feat/diagpart by @ChaosAbyssF in #25
• Feat/switch & Feat/merge by @ChaosAbyssF in #22
• add max and logicalnot op by @ssjcode in #29
• Dev/assign by @yaotianhang in #9
• Dev wukong by @wyfwyf-ux in #14
• fix: refactor mem allocator with BFC and optimize some kernel by @tngchien in #31
• feat: optimize BFC memory allocation and some kernel by @tngchien in #34
• feat: some improvment by @tngchien in #35
• Feat/reorg by @ChaosAbyssF in #36
• improve pack & unpack op performance by @greenhandF in #37
• feat: move pack location by @greenhandF in #38
• feat: isNanOp by @yaotianhang in #5
• feat: aloyshas pr containing conv2d op and expm1 op by @Aloyshaaaa in #33
• Feat/einsum by @ChaosAbyssF in #32
• fix: remove timing marco by @tngchien in #41
• feat: fix bug in resourceGather by @ssjcode in #43
• feat: replace musaMalloc with tf allocator in pack op by @greenhandF in #42
• Fix/reorg by @ChaosAbyssF in #40
• Fix/running device by @ChaosAbyssF in #44
• Dev/softplus by @yaotianhang in #30
• feat: switch to tf2.6 by @tngchien in #45
• feat: init fusion machanism by @tngchien in #56
• feat:timing_marco by @XFDG in #57
• Tf2.6 by @wyfwyf-ux in #58
• add new op TopK V2 by @yaotianhang in #64
• Dev/softplus by @yaotianhang in #63
• feat: add ReverseV2 musa kernel and tests by @Zhang-joey in #62
• add new op TensorListFromTensor by @yaotianhang in #61
• add new op TensoListReserve and Tensorlist utils(developed by Chaos) by @yaotianhang in #60
• Feat/IdentityN by @ChaosAbyssF in #59
• Aloyshaaaa by @Aloyshaaaa in #48
• add new op TensorListStack by @yaotianhang in #55
• Feat/assert by @ChaosAbyssF in #51
• Feat/all by @ChaosAbyssF in #50
• feat: add fused tokenmixer op by @greenhandF in #65
• Feat/matrix set diag v3 by @ChaosAbyssF in #54
• new fix pr by @Aloyshaaaa in #67
• Feat/EmptyTensorList by @ChaosAbyssF in #52
• Feat/where by @ChaosAbyssF in #49
• Fix/where by @ChaosAbyssF in #69
• feat: ReverseV2 by @Zhang-joey in #70
• aloysha by @Aloyshaaaa in #71
• fix: optimize zero copy for some kernel impl by @tngchien in #77
• CI Runner by @welo516 in #80
• Test by @welo516 in #76
• Add manual trigger by @welo516 in #81
• Tf2.6 by @wyfwyf-ux in #79
• Test add permisssion by @welo516 in #83
• solve bug by @welo516 in #91
• Feat tensordot fusion2 by @ssjcode in #92
• Only Update pr-validation.yml by @welo516 in #93
• Only Update pr-validation.yml by @welo516 in #94
• delete CI commit by @welo516 in #95
• feat:fused_gelu v1 by @XFDG in #72
• Opt/addn by @ChaosAbyssF in #73
• Fix ResourceApplyAdam resource variable update and session close crash by @XFDG in #78
• feat: Refactor MUSA device scheduling layer, fix thread starvation an… by @tngchien in #99
• fix: 增加daily测试/修改host端运行 by @welo516 in #100
• feat ：muDNNgelu + benchmark by @XFDG in #101
• Test permission issues by @welo516 in #107
• feat: add tensordot bias fusion by @ssjcode in #105
• feat: fused_clip and related test by @awexxxx in #103
• Feat/LinearReluFusion by @ChaosAbyssF in #102
• Feat/FusedSigmoidCalibration by @ChaosAbyssF in #104
• Hot/FusedSigmoidCalibration by @ChaosAbyssF in #108
• fix tensordot fusion op by @ssjcode in #110
• feat：Optimize AddV2 hot path and reduce profiling noiseAdd by @XFDG in #113
• add tensorflow musa playground by @real-zhangzhe in #112
• Add full-network performance testing by @welo516 in #111
• Newaloysha by @Aloyshaaaa in #106
• fix: fix mem leak by @tngchien in #120
• Fix applyadam by @ssjcode in #123
• feat: add PRelu fusion Op and improve cast and neg op by @greenhandF in #119
• feat: Optimize Select op by @awexxxx in #115
• feat: normalize fusion op by @ssjcode in #127
• add test count and introduce --fusion option by @welo516 in https://github.com/MooreThreads/tensorfl...