tvm
tvm copied to clipboard
apache
Open deep learning compiler stack for cpu, gpu and specialized accelerators
This PR implements quantization for hexagon and also adds schedule for nhwc-8h8w32c-2d layout. For this op, the input type is float32 and output can be uint8 or int8. cc @mehrdadh
For description see this [forum topic](https://discuss.tvm.apache.org/t/exporting-generated-functions-input-and-output-sizes-in-the-model-library-format/13532). Tagged reviewers: @areusch (don't know who else can review microTVM-related code)
[Bug] error: stack frame size (401472) exceeds limit (131056) in function 'default_function_kernel0'
Thanks for participating in the TVM community! We use https://discuss.tvm.ai for any general usage questions and discussions. The issue tracker is used for actionable items such as feature proposals discussion,...
hi, cc @AndrewZhaoLuo When I import my onnx model. I found my model get a error with pad. because the `pad` constant value is none. shown as follow: It`s just...
Refer to issue https://github.com/apache/tvm/issues/10211. The CSE pass can't handle commutativity because the arith system may not be able to do the commutativity。 The determination of the equality of two expressions...
Add string before doc, slow. Add string after doc, fast Thanks for contributing to TVM! Please refer to guideline https://tvm.apache.org/docs/contribute/ for useful information and tips. After the pull request is...
This PR fixes issue https://github.com/apache/tvm/issues/12763, where some OP are marked to keep the original dtype but some of its input is `bfloat16` while a `Cast` is missing. The AMP tests...
In meta schedule's featurization, arithmetic intensity was incorrectly calculated as log(FLOPs) / log(bytes). This change corrects it to log(FLOPs/bytes). Note that this is the same issue as in #12079. @vinx13...
TVM has recently switched to C++17. CUDA support with C++17 requires CMake past version 3.18, according to @vinx13. However, the current CMake version check in `CMakeLists.txt` is not checking for...
These currently take in the neighborhood of 30-40 minutes, we could reduce this time by - [ ] installing cmake binaries from a PPA instead of building from source (see...
