tvm
tvm copied to clipboard
apache
Open deep learning compiler stack for cpu, gpu and specialized accelerators
I am trying to compile TVM from the source in order to have microTVM support and GPU acceleration using Cuda and Tensorflow. ### Expected behavior TVM compiled from source, microTVM...
This pr fixes the tests that will be broken when we will update TFLite to the 2.9 version. We will update TensorFlow and TFLite versions to 2.9 so that we...
The patch including four parts: 1. Enable nn.contrib_dense_pack to be partitioned and offloaded by dnnl byoc. 1. Enable alter dense layout function during build relay model by introducing nn.contrib_dense_pack. 2....
After installing TVM correctly on this machine when trying to compile a model from Keras using `relay.frontend.from_keras(model, shape_dict)` defining target either as `target = tvm.target.cuda()` I get `[17:27:26] ../src/target/target_kind.cc:163: Warning:...
Continuation of https://github.com/apache/tvm/pull/10008 I was struggling to parse some relay graphs with custom data type(BYODT), generating parsing error. I found the [issue](https://github.com/apache/tvm/issues/8046) and above PR, tried to test it. All...
There are some frontend tests that fail on AArch64. They were disabled in prs: #12255, #12256 and #12257 List of failing tests: - tests/python/frontend/pytorch/test_forward.py::test_convert_torch_script_with_input_types - tests/python/frontend/onnx/test_forward.py::test_onnx_nodes[test_resize_upsample_sizes_nearest-llvm] - tests/python/frontend/tensorflow/test_forward.py::test_forward_crop_and_resize
It would be great if we can capture runtime in MLF since it is very useful specially for debugging auto tuning failures in microTVM auto tuning. cc @alanmacd @gromero
Current IntImm Node can't express negative number by hexadecimal. Eg: x = tvm.tir.const(0xFFFFFFFF, "int32") x.value should be -1 rather than 4294967295 Whether we need to consider this issue?
- add a utility function to help with debugging / investigation The function is particularly geared to help programmers examine changes to TVM modules and functions as they progress through...
### This issue is to track progress for C Devicea API - [x] https://github.com/apache/tvm-rfcs/pull/31
