tvm
tvm copied to clipboard
apache
Open deep learning compiler stack for cpu, gpu and specialized accelerators
This PR added a tuple-sum based implementation of layer norm. It performs one-pass reduction to compute mean and variance at the same time. Reducer pattern is also added to allow...
The logic in https://github.com/apache/tvm/blob/52d6b59a39f503fe382b4d7cbac4b02f9e44aae0/python/tvm/topi/generic/conv2d.py#L480-L499 is supposed to legalize the input dtype to be able to apply target-specific intrinsics that only support one of int8 or uint8. For example, the x86...
As discussed in https://github.com/apache/tvm-rfcs/blob/main/rfcs/0067-quarterly-releases.md, releases will happen roughly every three months. I have been put in charge for this one. This thread is meant to discuss the release schedule. We...
Add a new test case for torch `aten::fill_` operator implementation. This new test is testing a case of usage of a sequence of operators `aten::div()` and `aten::fill_() `that was causing...
This PR: - Refactors common project options to a single place for template projects - Adds a feature to update project options for customization - Adds a Makefile for Arduino...
This commit adds a new sub-analyzer, `TransitiveComparisonAnalyzer`, which attempts to apply multiple known comparisons to prove an unknown. For example, `a
Implement `Target::GetTargetDeviceType` (C++) or `get_target_device_type` (python) to get the device type (kDL...) for a given target. The attribute "target_device_type" can be used to override the default device type associated with...
This PR adds the option to save a generated micro project at a custom path, and to open a saved project and start communicating with it (instead of generating a...
The VM requires arguements to not be wrapped in an array. Passing the arguments unwrapped now. Also added relevant imports. @AndrewZhaoLuo
It appears the path has changed to the python github utils so I have refactored things. cc @driazati @areusch