Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

RuntimeError: Feature 'TVM_OP' is unknown, known features are: ['CUDA', 'CUDNN'........

See original GitHub issue

I was able to execute successfully last week and the environment didn’t change, but I suddenly reported an error today

The error information is as follows:

Traceback (most recent call last):
  File "example_advanced_tabular.py", line 6, in <module>
    import autogluon as ag
  File "/home/xxx/miniconda3/envs/mxnet-1.4/lib/python3.6/site-packages/autogluon/__init__.py", line 6, in <module>
    from .utils.try_import import *
  File "/home/xxx/miniconda3/envs/mxnet-1.4/lib/python3.6/site-packages/autogluon/utils/__init__.py", line 5, in <module>
    from .dataset import *
  File "/home/xxx/miniconda3/envs/mxnet-1.4/lib/python3.6/site-packages/autogluon/utils/dataset.py", line 3, in <module>
    import mxnet as mx
  File "/home/xxx/miniconda3/envs/mxnet-1.4/lib/python3.6/site-packages/mxnet/__init__.py", line 90, in <module>
    from . import tvmop
  File "/home/xxx/miniconda3/envs/mxnet-1.4/lib/python3.6/site-packages/mxnet/tvmop.py", line 22, in <module>
    if Features().is_enabled("TVM_OP"):
  File "/home/xxx/miniconda3/envs/mxnet-1.4/lib/python3.6/site-packages/mxnet/runtime.py", line 93, in is_enabled
    feature_name, list(self.keys())))
RuntimeError: Feature 'TVM_OP' is unknown, known features are: ['CUDA', 'CUDNN', 'NCCL', 'CUDA_RTC', 'TENSORRT', 'CPU_SSE', 'CPU_SSE2', 'CPU_SSE3', 'CPU_SSE4_1', 'CPU_SSE4_2', 'CPU_SSE4A', 'CPU_AVX', 'CPU_AVX2', 'OPENMP', 'SSE', 'F16C', 'JEMALLOC', 'BLAS_OPEN', 'BLAS_ATLAS', 'BLAS_MKL', 'BLAS_APPLE', 'LAPACK', 'MKLDNN', 'OPENCV', 'CAFFE', 'PROFILER', 'DIST_KVSTORE', 'CXX14', 'INT64_TENSOR_SIZE', 'SIGNAL_HANDLER', 'DEBUG']

operating environment

python 3.6
mxnet-cu90
cuda 9.0

Issue Analytics

State:
Created 3 years ago
Comments:7 (1 by maintainers)

Top GitHub Comments

1reaction

zzdgitcommented, Apr 18, 2020

I tried version 1.6 of mxnet，but still reported an error。Then I switched to a server and it was fine。 I wonder what dependencies need to be installed to support TVM_OP。 thanks,

0reactions

Innixmacommented, Apr 22, 2020

@hanyunfan If the solution mentioned here did not work for you, please feel free to create a new issue highlighting your situation.

Top Results From Across the Web

RuntimeError: Feature 'TVM_OP' is unknown, known ... - GitHub

I was able to execute successfully last week and the environment didn't change, but I suddenly reported an error today The error information ......

Failed to get convolution algorithm. This is probably because ...

If using Conda environments, in my case the issue was solved by installing tensorflow-gpu and not CUDAtoolkit nor cuDNN because they are already ......

RuntimeError: cuDNN error - PyTorch Forums

I suppose the problem comes from the fact that I'm trying to use torch with CUDA 11.1 with an incompatible Driver Version ,...

CUDA Toolkit 12.0 Released for General Availability

CUDA 12.0 exposes programmable functionality for many features of the NVIDIA Hopper and NVIDIA Ada Lovelace architectures:.