Naren Dasan
Naren Dasan
# Description Small changes to address outstanding issues Fixes #266 Fixes #834 ## Type of change Please delete options that are not relevant and/or add your own. - Bug fix...
https://github.com/pytorch/TensorRT/blob/5ad982639c68dbc079b211ad3ffef6ec900b0dcc/core/conversion/converters/impl/element_wise.cpp#L28
**Is your feature request related to a problem? Please describe.** Right now we require a specific converter definition for each operator we support. In the evaluator library we support the...
**Is your feature request related to a problem? Please describe.** Currently TRTorch can only return programs that can be represented as ``` f(Tensor[]) -> Tensor[] ``` So this includes cases...
**Is your feature request related to a problem? Please describe.** Notebooks right now have a few issues including: - Notebooks must be run AOT to avoid triggering execution during documentation...
In order to fully support RNNs, we want to be able to make recurrent subgraphs to TRT. To do so we must expand the capability of the compiler to recognize...
## Bug Description When using dynamic batch, `aten::add` dimension checks fail reporting: ``` ERROR: [TRTorch Conversion Context] - %457 : Tensor = aten::add_(%456, %455, %3): broadcast dimensions must be non-negative...
It would be nice to be able to click on a figure to get an expanded view like you could do on the New York Times site (e.g. https://www.nytimes.com/2017/10/16/dining/kimbal-musk-food.html). Having...
Seems like the output buffer order is alphabetical but we should have a way to record / query what the order in the TensorList would be so we dont need...
We should move to using platforms (https://docs.bazel.build/versions/master/platforms.html) and toolchains (https://docs.bazel.build/versions/master/toolchains.html) as soon as the rules are stable so that we can version the toolchains between PDK versions. A hacky version...