1 Release published by 1 person

• v2.58.0 Release 2.58.0 corresponding to NGC container 25.05

  published May 31, 2025

3 Pull requests merged by 2 people

• fix: Improve cancellation handling for gRPC non-decoupled inference

  #8220 merged May 29, 2025

• fix: fix the config.pbtxt example after converting the tensorflow to onnx

  #8223 merged May 28, 2025

• fix: fix the L0_infer tests for expected num tests

  #8222 merged May 27, 2025

3 Pull requests opened by 3 people

• Support max_completion_tokens option in OpenAI frontend

  #8226 opened May 30, 2025

• 25.05 release notes

  #8227 opened May 30, 2025

• docs: update the link formats for additional security networking guides

  #8229 opened Jun 2, 2025

3 Issues opened by 3 people

• is there any examples or tutorials for tensorrt backend Optimization?

  #8228 opened Jun 1, 2025

• Unable to use packages installed in env.tar.gz in later versions of triton

  #8225 opened May 30, 2025

• Question about the Future of "PyTorch 2.0 Backend [Experimental]" in Triton Server Python (LibTorch) Backend

  #8224 opened May 30, 2025

5 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• fix: Update handling of large array sizes

  #8174 commented on Jun 2, 2025 • 18 new comments

• [Question] Reloading models and latency spikes

  #8117 commented on May 27, 2025 • 0 new comments

• AWS ECR identifies vulnerability with Pillow version for Triton server image 23.07

  #6222 commented on May 29, 2025 • 0 new comments

• What's the difference when starting tritonserver with `mpirun --allow-run-as-root -n 1 /opt/tritonserver/bin/tritonserver` vs. `/opt/tritonserver/bin/tritonserver` directly?

  #7371 commented on Jun 1, 2025 • 0 new comments

• Histogram Metric for multi-instance tail latency aggregation

  #7672 commented on Jun 2, 2025 • 0 new comments