tvm
tvm copied to clipboard
apache
Open deep learning compiler stack for cpu, gpu and specialized accelerators
As per title. [torchvision.ops.stochastic_depth](https://pytorch.org/vision/main/generated/torchvision.ops.stochastic_depth.html) isn't a pytorch core ops, but some pytorch image classification models such as convnext, inception_v3, and maxvit_t are using it. So I think it's worth supporting...
As per title. I'll leave mobilenet_v2 part as it is but we could remove keras dependency since torchvision provides a mobilenet_v2 pretrained model. Not sure which is better. Related issue...
Replace mxnet's resnet18 with torchvision. Clean up import statements in `rust/tvm/examples/resnet/src/build_resnet.py`. Related PR and Issue: - #16546 - #16572 - #16547 cc: @Hzfengsy @tqchen @cgerum
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,...
### Actual behavior ``` Traceback (most recent call last): File "/share_container/optfuzz/res/bugs/assert_lazy.py", line 52, in mod = relax.transform.LazyTransformParams()(mod) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/software/tvm-lunder/python/tvm/ir/transform.py", line 238, in __call__ return _ffi_transform_api.RunPass(self, mod) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/software/tvm-lunder/python/tvm/_ffi/_ctypes/packed_func.py",...
### Actual behavior ``` Traceback (most recent call last): File "/share_container/optfuzz/res/bugs/inconsis222.py", line 258, in np.testing.assert_allclose(before_outputs, after_outputs, 1e-3, 1e-3) File "/root/miniconda3/lib/python3.12/site-packages/numpy/testing/_private/utils.py", line 1504, in assert_allclose assert_array_compare(compare, actual, desired, err_msg=str(err_msg), File "/root/miniconda3/lib/python3.12/contextlib.py",...
These tests were found to be flaky (intermittently failing on `main` or failed in a PR with unrelated changes). See [the docs](https://github.com/apache/tvm/blob/main/docs/contribute/ci.rst#handling-flaky-failures) for details. ### Tests(s) - `tests/python/autotvm/test_autotvm_droplet_tuner.py::test_multi_filter` ### Jenkins...
Hi all, The pass `RemoveUnusedOutputs` seems to give an unexpected optimized result. Due to the lack of detailed documentation about this API (e.g., `relax.transform.RemoveUnusedOutputs`), I cannot confirm if the optimization...
In general, intermediate values produced while evaluating Relax functions are not visible to an end user. While this provides stronger guarantees to the compiler, it can make debugging difficult. For...
error from running the below script ``` Binary was created using {clml} but a loader of that name is not registered. Available loaders are opencl, GraphRuntimeFactory, GraphExecutorFactory, static_library, relax.Executable, AotExecutorFactory,...
