(Disclaimer: This is a duplicate of the issue in TIS Github #8154 as there was no response since it was posted 2 weeks ago)
Given a model with multiple versions, I have written some logic to load a specific version of a model using EXPLICIT mode by modifying the runtime model config (see #8153 for details).
Could there be a simpler way?
BR,
Scot.
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| How to get model configuration from HTTP API without first loading the model in EXPLICIT mode? | 1 | 32 | April 30, 2025 | |
| Wrong model versioning when using gRPC | 9 | 850 | August 16, 2022 | |
| Using TLT models with Triton Inference Server | 8 | 1467 | October 12, 2021 | |
| Loading a new plan file while running inference | 1 | 623 | June 9, 2020 | |
| Triton deployment using Explicit model control | 1 | 1092 | October 14, 2021 | |
| Bug : Tensorrt Model not loading on same GPU on a different device (slight driver version difference) | 1 | 248 | April 30, 2024 | |
| Cannot re-use TensorRT models between docker container and on-device | 7 | 39 | September 3, 2025 | |
| TensorRT Model load error in Deepstream v6.0.1 with TensorRT 8.5.3.1 | 9 | 1500 | February 10, 2023 | |
| Convert pb saved model obtained from Tensorflow Object detection API to a TensorRT model and load it to TensorRT Inference Server | 1 | 597 | October 25, 2018 | |
| Riva_server fails if not all Triton models are loaded | 6 | 1102 | January 27, 2023 |