I’m attempting to run the from_tensorflow.py example (https://tvm.apache.org/docs/tutorials/frontend/from_tensorflow.html) and compare NCHW to NHWC by changing the format accordingly. However, I am getting the errors “cannot find config for conv2d_nchw.cuda”, “cannot find conv2d_nhwc.cuda”, and “cannot find config for dense_small_batch”. I’ve read that I need to autotune them but which source code do I use? I’ve already tried “Auto-tuning a convolutional network for NVIDIA GPU” (https://tvm.apache.org/docs/tutorials/autotvm/tune_relay_cuda.html) but it didn’t seem to create configurations for conv2d_nchw.cuda and the others.
