Apurba Bose
Apurba Bose
Hi @narendasan I tested this locally and the test passes now.
I am not sure exactly, but should not these test cases be included in the test_dynamic_fallback.cpp in the TensorRT/tests/cpp tests?
Closing the issue as it has been tested.
PR #2740
PR #2519
The `harness.py` file is changed in this for the cases in which the index is passed in the `forward` function. torch scatter function takes only int64 inputs for index Couple...
Thanks for pointing this out and the above changes! Couple of points- 1. `grid_sampler_2d `would require some additional changes 2. As pointed out by you (thanks!), `upsample` cases do not...
Looks good to me, pending on CI
Had a doubt on this PR. Suppose we feed in int64 ITensor for the test case directly as below, the TRTInterpretor would complain saying int64 inputs are not allowed. Meanwhile...
@gs-olive int32 in the test case leads to this error- RuntimeError: gather(): Expected dtype int64 for index