1 Release published by 1 person

• v0.17.0.rc0 Apache TVM v0.17.0.rc0

  published Jul 20, 2024

17 Pull requests merged by 10 people

• Remove and replace deprecated distutils.util.strtobool()

  #17185 merged Jul 23, 2024

• [DLIGHT][GPU] Add OpenCL dequant matmul schedule

  #17187 merged Jul 23, 2024

• [Disco] Group-wise operation

  #17180 merged Jul 23, 2024

• [MetaSchedule] Replace xgboost.rabit with xgboost.collective because it's deprecated

  #17166 merged Jul 23, 2024

• [Relax][PyTorch] Add support for torch.einsum

  #17186 merged Jul 23, 2024

• Add packaging to python/gen_requirements.py

  #17188 merged Jul 22, 2024

• [Hexagon] [CMake] Fix v66 build issue

  #17169 merged Jul 22, 2024

• [FFI] Add python signal handler for ctypes FFI

  #17181 merged Jul 22, 2024

• [Relax][PyTorch] Add support for torch.permute

  #17184 merged Jul 22, 2024

• [Relax] Integrate cuDNN attention

  #17157 merged Jul 22, 2024

• [MetaSchedule]Add a testcase for padded conv2d in meta_schedule

  #17171 merged Jul 22, 2024

• [Relay][FQ2I]: Use appropriate dtype while quantizing relay.op.nn.pad…

  #17177 merged Jul 22, 2024

• Use packaging.version.parse instead of distutils.version.LooseVersion

  #17173 merged Jul 19, 2024

• [TVMJS] Check DataType.NUMPY2STR when saving array

  #17174 merged Jul 18, 2024

• [Relax] [ONNX] Add support for Sign and Not

  #17167 merged Jul 18, 2024

• [Meta Schedule][XGBoost] enable custom callback func test with xgboost>=1.6.0

  #17168 merged Jul 18, 2024

• [Relax][BugFix] Fix a bug about the IR construction in test file

  #17121 merged Jul 18, 2024

6 Pull requests opened by 6 people

• Pass to eliminate redundant branch and overcompute

  #17170 opened Jul 18, 2024

• [Disco] Implement SocketSession

  #17182 opened Jul 21, 2024

• Add support for `torch.nn.functional.max_pool2d`

  #17189 opened Jul 23, 2024

• [Cython][FFI] Fix crash when call del operator for handle

  #17190 opened Jul 23, 2024

• [Disco] Cross-group and p2p send/receive primitives

  #17191 opened Jul 23, 2024

• [KVCache] Partial layers support

  #17192 opened Jul 23, 2024

1 Issue closed by 1 person

• [Bug] pytorch relax frontend failed to import models with torch.permute

  #17183 closed Jul 22, 2024

5 Issues opened by 3 people

• [VOTE] Release Apache TVM v0.17.0.rc0

  #17179 opened Jul 20, 2024

• [Release] v0.17.0 Release Candidate Notes

  #17178 opened Jul 20, 2024

• [Relax][Bug] Cannot find PackedFunc tir_zeros

  #17176 opened Jul 18, 2024

• [Relax][Bug] CodeGenVM cannot handle this intrinsic tensor_to_shape

  #17175 opened Jul 18, 2024

• [Tracking Issue] Remove `distutils` dependency from the codebase

  #17172 opened Jul 18, 2024

9 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• [TIR][Analyzer] Simplify `x==x` expressions for all dtypes

  #17158 commented on Jul 19, 2024 • 2 new comments

• [Bug] Building for qualcomm hexagon dsp V66 architeture

  #17163 commented on Jul 18, 2024 • 0 new comments

• [BUG]: Relay.transform.FoldScaleAxis() Error in folding the SqueezeExcitation structure of mobilenetv3.

  #17151 commented on Jul 20, 2024 • 0 new comments

• [Release] v0.17.0 release schedule

  #17122 commented on Jul 20, 2024 • 0 new comments

• [Bug] MicroTVM Compilation fails for ARM Cortex-M4 ,with CMSIS-NN=false

  #17136 commented on Jul 23, 2024 • 0 new comments

• [FFI][RUNTIME] Introduce runtime boxed types for int/float/bool

  #16183 commented on Jul 18, 2024 • 0 new comments

• [CLML][CI] Fix for few clml regression issues

  #17117 commented on Jul 22, 2024 • 0 new comments

• [Relax] Implement R.ensure_zero_offset and update memory planning for R.view

  #17145 commented on Jul 17, 2024 • 0 new comments

• [KVCache] PagedKVCache Quantization

  #17159 commented on Jul 19, 2024 • 0 new comments