Tamas Bela Feher
Tamas Bela Feher
No, TensorRT is not only computer vision oriented. We have an [RNN code sample](https://github.com/NVIDIA/TensorRT/tree/main/samples/sampleCharRNN) which actually shows how to define an LSTM.
Thanks for the report. It should not try to rebuild the engines again, but it is difficult to say anything more at this stage. It might help if you can...
The problem still exist. Here is a modified reproducer, which adds a TRT incompatible op (DataFormatVecPermute) followed by other TRT compatible ops: ```python #!/usr/bin/env python # coding: utf-8 import os...
- As @mankeyboy said, using `minimum_segment_size=3` or larger should help to get around the conversion problem. - The segfault which is present in NGC 20.01-tf2-py3 is not there in the...
Hi, there was recently a change that disabled TRT conversion of transpose ops on tensors that have more than 2.5 million elements. Normally this would cause the network to be...
Hi, it looks like the same issue as #181. That one is fixed in TF master, please give it a try.
Suggestions for improved naming - [ ] benchmark name (in build.sh) https://github.com/rapidsai/cuvs/pull/130#discussion_r1619512483 - [ ] name of index class (in ann_type.hpp) https://github.com/rapidsai/cuvs/pull/130#discussion_r1619515306
cross referencing the FAISS issue where the problem was encountered https://github.com/facebookresearch/faiss/issues/3621
/ok to test
/ok to test