# \#tensorrt

**URL:** https://forums.developer.nvidia.com/tag/tensorrt/12.md

[Latest](https://forums.developer.nvidia.com/latest.md) · [Categories](https://forums.developer.nvidia.com/categories.md) · [Tags](https://forums.developer.nvidia.com/tags.md)

---

## [One Orin Nano Super, one camera, no cloud AI: meet Sentinel (open source, testers wanted)](https://forums.developer.nvidia.com/t/one-orin-nano-super-one-camera-no-cloud-ai-meet-sentinel-open-source-testers-wanted/389377)

<div class="topic-metadata">

**Author:** [@hemkesh.2021](https://forums.developer.nvidia.com/u/hemkesh.2021)\
**Replies:** 2\
**Last updated:** [November 18, 2026, 7:03am UTC](https://forums.developer.nvidia.com/t/one-orin-nano-super-one-camera-no-cloud-ai-meet-sentinel-open-source-testers-wanted/389377 "2026-11-18T07:03:35Z")

</div>

One Orin Nano Super. One RTSP camera. Every model runs on the board. Sentinel spots people, follows them across the frame, and opens an incident when someone steps into a zone you care about. Then it tells you, even if …

---

## [Orin Nano Super, JetPack 6.2.1: CPU at 95 % while GPU clock stays below max during TensorRT inference on 4 RTSP streams](https://forums.developer.nvidia.com/t/orin-nano-super-jetpack-6-2-1-cpu-at-95-while-gpu-clock-stays-below-max-during-tensorrt-inference-on-4-rtsp-streams/388647)

<div class="topic-metadata">

**Author:** [@vishalan](https://forums.developer.nvidia.com/u/vishalan)\
**Replies:** 2\
**Last updated:** [November 18, 2026, 6:49am UTC](https://forums.developer.nvidia.com/t/orin-nano-super-jetpack-6-2-1-cpu-at-95-while-gpu-clock-stays-below-max-during-tensorrt-inference-on-4-rtsp-streams/388647 "2026-11-18T06:49:51Z")

</div>

Hi there, Need some assistance in diagnosing performance capabilities, to find out if hardware or software issue. The below indicates the setup and some performance observations: | SETUP | Device: Jetson Orin Nano 8GB…

---

## [Leap Motion Controller hand tracking on AGX Xavier: TensorRT FP16, about 80 fps, no Ultraleap software](https://forums.developer.nvidia.com/t/leap-motion-controller-hand-tracking-on-agx-xavier-tensorrt-fp16-about-80-fps-no-ultraleap-software/389138)

<div class="topic-metadata">

**Author:** [@angel.marquezaguilar](https://forums.developer.nvidia.com/u/angel.marquezaguilar)\
**Replies:** 1\
**Last updated:** [November 16, 2026, 7:30am UTC](https://forums.developer.nvidia.com/t/leap-motion-controller-hand-tracking-on-agx-xavier-tensorrt-fp16-about-80-fps-no-ultraleap-software/389138 "2026-11-16T07:30:22Z")

</div>

I got the original Leap Motion Controller tracking hands in 3D on a Jetson AGX Xavier (JetPack 5.1, TensorRT 8.5). The controller is read as a plain UVC camera, and Monado’s Mercury hand models run as FP16 TensorRT engin…

---

## [EfficientDetD4 ONNX conversion to TRT engine fails on ORIN NX SUPER 8GB](https://forums.developer.nvidia.com/t/efficientdetd4-onnx-conversion-to-trt-engine-fails-on-orin-nx-super-8gb/386017)

<div class="topic-metadata">

**Author:** [@adit\_bhrgv](https://forums.developer.nvidia.com/u/adit_bhrgv)\
**Replies:** 9\
**Last updated:** [November 12, 2026, 8:08am UTC](https://forums.developer.nvidia.com/t/efficientdetd4-onnx-conversion-to-trt-engine-fails-on-orin-nx-super-8gb/386017 "2026-11-12T08:08:31Z")

</div>

Description A clear and concise description of the bug or issue. Environment TensorRT Version: 10.3.0 GPU Type: Orin NX iGPU, ORIN NX SUPER 8GB Model Nvidia Driver Version: 540.4.0 CUDA Version: 12.6 CUDNN Version:…

---

## [Nline Logit-Level Drift Telemetry & Early Stopping Hooks in TensorRT-LLM (70–80% Power Reduction)](https://forums.developer.nvidia.com/t/nline-logit-level-drift-telemetry-early-stopping-hooks-in-tensorrt-llm-70-80-power-reduction/388814)

<div class="topic-metadata">

**Author:** [@dattathreya](https://forums.developer.nvidia.com/u/dattathreya)\
**Replies:** 0\
**Last updated:** [November 12, 2026, 6:33am UTC](https://forums.developer.nvidia.com/t/nline-logit-level-drift-telemetry-early-stopping-hooks-in-tensorrt-llm-70-80-power-reduction/388814 "2026-11-12T06:33:12Z")

</div>

Description Hi NVIDIA TensorRT Team & Community, We are developing Buddhi\_Core (Nivedita AI LLP, Patent App. 202641098749), an inline runtime inference governance and silicon power-optimization engine for LLMs and edge …

---

## [How can we run inference of Alpamayo 2 super on jetson AGX thor?](https://forums.developer.nvidia.com/t/how-can-we-run-inference-of-alpamayo-2-super-on-jetson-agx-thor/388740)

<div class="topic-metadata">

**Author:** [@neethujoseph02](https://forums.developer.nvidia.com/u/neethujoseph02)\
**Replies:** 1\
**Last updated:** [November 11, 2026, 7:25am UTC](https://forums.developer.nvidia.com/t/how-can-we-run-inference-of-alpamayo-2-super-on-jetson-agx-thor/388740 "2026-11-11T07:25:01Z")

</div>

how can we run inference of Alpamayo 2 super on jetson thor? What are the end to end steps?

---

## [Jetson Orin Nano TRT 10.16 — INT8 DETR engine slower than FP16; NaNs when decoder is FP16](https://forums.developer.nvidia.com/t/jetson-orin-nano-trt-10-16-int8-detr-engine-slower-than-fp16-nans-when-decoder-is-fp16/387395)

<div class="topic-metadata">

**Author:** [@jagdish.k](https://forums.developer.nvidia.com/u/jagdish.k)\
**Replies:** 3\
**Last updated:** [November 11, 2026, 12:49am UTC](https://forums.developer.nvidia.com/t/jetson-orin-nano-trt-10-16-int8-detr-engine-slower-than-fp16-nans-when-decoder-is-fp16/387395 "2026-11-11T00:49:49Z")

</div>

Hi, I am quantizing transformer detectors with TensorRT 10.16.2.10 on a Jetson Orin Nano (JetPack / L4T with CUDA, no DLA on this SKU). Power mode is 15W (GPU ~612 MHz). Builder: optimization level 3, workspace ~256 MB,…

---

## [DeepStream 6.4 pipeline latency build-up during multi-stream RTSP anti-tailgating inference on RTX 4000 Ada](https://forums.developer.nvidia.com/t/deepstream-6-4-pipeline-latency-build-up-during-multi-stream-rtsp-anti-tailgating-inference-on-rtx-4000-ada/387425)

<div class="topic-metadata">

**Author:** [@words7243](https://forums.developer.nvidia.com/u/words7243)\
**Replies:** 2\
**Last updated:** [November 3, 2026, 8:27am UTC](https://forums.developer.nvidia.com/t/deepstream-6-4-pipeline-latency-build-up-during-multi-stream-rtsp-anti-tailgating-inference-on-rtx-4000-ada/387425 "2026-11-03T08:27:46Z")

</div>

Hi All, We are deploying an edge vision analytics worker in a Tier-3 facility white space to monitor server pod corridors and airlock mantrap portals. As part of integrating automated anti-tailgating detection with our…

---

## [Orion - ROS2 Quadruped Robotics Platform](https://forums.developer.nvidia.com/t/orion-ros2-quadruped-robotics-platform/382473)

<div class="topic-metadata">

**Author:** [@shinyflame007](https://forums.developer.nvidia.com/u/shinyflame007)\
**Replies:** 1\
**Last updated:** [October 28, 2026, 6:06am UTC](https://forums.developer.nvidia.com/t/orion-ros2-quadruped-robotics-platform/382473 "2026-10-28T06:06:49Z")

</div>

Meet Orion - a ROS2-based quadruped robotics system I designed, built, and programmed from scratch over the course of 8 months. After countless nights of debugging and iterating, I’m thrilled to finally share it in actio…

---

## [FPS Drop / Performance Degradation from ~10 FPS to ~8 FPS While Parsing Camera RTSP Frames After OS, CUDA and Kubernetes Upgrade](https://forums.developer.nvidia.com/t/fps-drop-performance-degradation-from-10-fps-to-8-fps-while-parsing-camera-rtsp-frames-after-os-cuda-and-kubernetes-upgrade/381902)

<div class="topic-metadata">

**Author:** [@amit.k.jain](https://forums.developer.nvidia.com/u/amit.k.jain)\
**Replies:** 4\
**Last updated:** [October 19, 2026, 6:07am UTC](https://forums.developer.nvidia.com/t/fps-drop-performance-degradation-from-10-fps-to-8-fps-while-parsing-camera-rtsp-frames-after-os-cuda-and-kubernetes-upgrade/381902 "2026-10-19T06:07:50Z")

</div>

Please provide complete information as applicable to your setup. • Hardware Platform (GPU):- T4 • DeepStream Version:- 6.0 • JetPack Version (valid for Jetson only):- NA • TensorRT Version:- 8.0.1 • NVIDIA GPU Drive…

---

## [Fine-tuned DeepVariant model in Parabricks 4.7.1: converter, make\_examples version, 6-channel](https://forums.developer.nvidia.com/t/fine-tuned-deepvariant-model-in-parabricks-4-7-1-converter-make-examples-version-6-channel/385326)

<div class="topic-metadata">

**Author:** [@tealeave](https://forums.developer.nvidia.com/u/tealeave)\
**Replies:** 2\
**Last updated:** [October 8, 2026, 10:05pm UTC](https://forums.developer.nvidia.com/t/fine-tuned-deepvariant-model-in-parabricks-4-7-1-converter-make-examples-version-6-channel/385326 "2026-10-08T22:05:00Z")

</div>

Fine-tuned DeepVariant model in Parabricks 4.7.1: converter, make\_examples version, 6-channel From thread 315110 we understand, we fine-tune DeepVariant on plasma cell-free DNA (short fragments) with matched buffy-coat …

---

## [PyTorch Conditionals that run on GPU / TensorRT](https://forums.developer.nvidia.com/t/pytorch-conditionals-that-run-on-gpu-tensorrt/379608)

<div class="topic-metadata">

**Author:** [@jetson.lover](https://forums.developer.nvidia.com/u/jetson.lover)\
**Replies:** 4\
**Last updated:** [October 4, 2026, 10:23am UTC](https://forums.developer.nvidia.com/t/pytorch-conditionals-that-run-on-gpu-tensorrt/379608 "2026-10-04T10:23:01Z")

</div>

Description The TensorRT documentation for Working with Conditionals has an example for PyTorch. However, this relies on a Python conditional if c % 2 == 0: which runs on the CPU not GPU: def sum\_even(items): s = to…

---

## [Jetson Orin NX 8GB hard reboots during TensorRT engine build on fresh JetPack 7.2.1 / DeepStream 9.1](https://forums.developer.nvidia.com/t/jetson-orin-nx-8gb-hard-reboots-during-tensorrt-engine-build-on-fresh-jetpack-7-2-1-deepstream-9-1/382905)

<div class="topic-metadata">

**Author:** [@jamesebrogan](https://forums.developer.nvidia.com/u/jamesebrogan)\
**Replies:** 9\
**Last updated:** [October 2, 2026, 9:01pm UTC](https://forums.developer.nvidia.com/t/jetson-orin-nx-8gb-hard-reboots-during-tensorrt-engine-build-on-fresh-jetpack-7-2-1-deepstream-9-1/382905 "2026-10-02T21:01:02Z")

</div>

I have a Jetson Orin NX 8GB that reproducibly hard reboots, with no kernel panic or orderly shutdown visible, while TensorRT is building an engine. orin-nx-jp721-tensorrt-reboot-diagnostics.tar.gz (21.6 KB) I have repr…

---

## [Qualifying a DeepStream/TensorRT pipeline from x86 to Jetson Orin NX](https://forums.developer.nvidia.com/t/qualifying-a-deepstream-tensorrt-pipeline-from-x86-to-jetson-orin-nx/384748)

<div class="topic-metadata">

**Author:** [@garygayle](https://forums.developer.nvidia.com/u/garygayle)\
**Replies:** 1\
**Last updated:** [October 1, 2026, 4:29am UTC](https://forums.developer.nvidia.com/t/qualifying-a-deepstream-tensorrt-pipeline-from-x86-to-jetson-orin-nx/384748 "2026-10-01T04:29:05Z")

</div>

We’ve completed an x86 DeepStream/TensorRT qualification path using a frozen YOLOv8m ONNX authority and are preparing the next qualification step on Jetson Orin NX. Current validated gates include: frozen model lineag…

---

## [Where is TensorRT on Jetson Orin Nano in Jetpack 7.2.1?](https://forums.developer.nvidia.com/t/where-is-tensorrt-on-jetson-orin-nano-in-jetpack-7-2-1/384616)

<div class="topic-metadata">

**Author:** [@jetson.lover](https://forums.developer.nvidia.com/u/jetson.lover)\
**Replies:** 3\
**Last updated:** [September 30, 2026, 3:01am UTC](https://forums.developer.nvidia.com/t/where-is-tensorrt-on-jetson-orin-nano-in-jetpack-7-2-1/384616 "2026-09-30T03:01:45Z")

</div>

Description According to the Jetpack 7.2.1 iso download page, TensorRT 10.16.2 is supported by default. However, running the commands dpkg -l | grep nvinfer does not show any output, that hints that TensorRT is actua…

---

## [Wrong FP16 results when a fused convolution, activation and residual add reads a strided concat view](https://forums.developer.nvidia.com/t/wrong-fp16-results-when-a-fused-convolution-activation-and-residual-add-reads-a-strided-concat-view/383599)

<div class="topic-metadata">

**Author:** [@Y-T-G](https://forums.developer.nvidia.com/u/Y-T-G)\
**Replies:** 0\
**Last updated:** [September 18, 2026, 9:17am UTC](https://forums.developer.nvidia.com/t/wrong-fp16-results-when-a-fused-convolution-activation-and-residual-add-reads-a-strided-concat-view/383599 "2026-09-18T09:17:16Z")

</div>

Description On TensorRT 11.2.1.2 a fused FP16 region of the form convolution, activation, residual add returns wrong values when the add operand is a non dense view of a later concatenation buffer. The fused layer reads…

---

## [Alpamayo-R1-10B TensorRT engine OOM on DRIVE AGX Thor](https://forums.developer.nvidia.com/t/alpamayo-r1-10b-tensorrt-engine-oom-on-drive-agx-thor/382500)

<div class="topic-metadata">

**Author:** [@nandhini.ravikumar1](https://forums.developer.nvidia.com/u/nandhini.ravikumar1)\
**Replies:** 1\
**Last updated:** [September 13, 2026, 7:08am UTC](https://forums.developer.nvidia.com/t/alpamayo-r1-10b-tensorrt-engine-oom-on-drive-agx-thor/382500 "2026-09-13T07:08:12Z")

</div>

I am deploying nvidia/Alpamayo-R1-10B with TensorRT-Edge-LLM on DRIVE AGX Thor running DriveOS 7.2.5.0. CUDA: 13.3 TensorRT reported by llm\_build: 11.0.1 SDK container: driveos-sdk … 7.2.5.0-0004 Target: auto-thor T…

---

## [ROS 2 Autonomous Mobile Robot with TensorRT Vision, LiDAR and Voice AI on Jetson Orin Nano](https://forums.developer.nvidia.com/t/ros-2-autonomous-mobile-robot-with-tensorrt-vision-lidar-and-voice-ai-on-jetson-orin-nano/382189)

<div class="topic-metadata">

**Author:** [@sourabhwarghane](https://forums.developer.nvidia.com/u/sourabhwarghane)\
**Replies:** 2\
**Last updated:** [September 3, 2026, 5:11pm UTC](https://forums.developer.nvidia.com/t/ros-2-autonomous-mobile-robot-with-tensorrt-vision-lidar-and-voice-ai-on-jetson-orin-nano/382189 "2026-09-03T17:11:03Z")

</div>

Hi NVIDIA Jetson Community, I’m excited to share my Autonomous Mobile Robot (AMR) built around the NVIDIA Jetson Orin Nano 8GB and ROS 2 Humble. I started this project while learning ROS 2 and AI robotics, and graduall…

---

## [Supporting TRT 11.2.1 on Jetpack or Drive OS](https://forums.developer.nvidia.com/t/supporting-trt-11-2-1-on-jetpack-or-drive-os/381922)

<div class="topic-metadata">

**Author:** [@mac1234](https://forums.developer.nvidia.com/u/mac1234)\
**Replies:** 3\
**Last updated:** [September 2, 2026, 6:34am UTC](https://forums.developer.nvidia.com/t/supporting-trt-11-2-1-on-jetpack-or-drive-os/381922 "2026-09-02T06:34:51Z")

</div>

Are there plans to support TRT version 11.2.1 on upcoming Jetpack or DriveOS releases officially ? Thanks

---

## [Skills and DLI course path for a 4-camera DeepStream detection pipeline — what are we missing?](https://forums.developer.nvidia.com/t/skills-and-dli-course-path-for-a-4-camera-deepstream-detection-pipeline-what-are-we-missing/381195)

<div class="topic-metadata">

**Author:** [@maimoon](https://forums.developer.nvidia.com/u/maimoon)\
**Replies:** 5\
**Last updated:** [August 31, 2026, 9:51am UTC](https://forums.developer.nvidia.com/t/skills-and-dli-course-path-for-a-4-camera-deepstream-detection-pipeline-what-are-we-missing/381195 "2026-08-31T09:51:23Z")

</div>

Hello, We are building an edge vision product on Jetson Orin NX and are putting together the capability plan for a small team. I would value a sanity check on what we have identified, and guidance on which DLI courses m…

---

## [Endoscopic video SR at 60 fps with \<=8 ms p99 12 ms: Holoscan or direct CUDA?](https://forums.developer.nvidia.com/t/endoscopic-video-sr-at-60-fps-with-8-ms-p99-12-ms-holoscan-or-direct-cuda/381670)

<div class="topic-metadata">

**Author:** [@john.chen7](https://forums.developer.nvidia.com/u/john.chen7)\
**Replies:** 0\
**Last updated:** [August 28, 2026, 7:48am UTC](https://forums.developer.nvidia.com/t/endoscopic-video-sr-at-60-fps-with-8-ms-p99-12-ms-holoscan-or-direct-cuda/381670 "2026-08-28T07:48:06Z")

</div>

Context We are a medical device company with a cleared electronic endoscopic vision platform in the field, and an NVIDIA Inception member. We are adding real-time 2x super-resolution to the live surgical video path in t…

---

## [Onnx to TRT conversion doesn't work on Orin Nano](https://forums.developer.nvidia.com/t/onnx-to-trt-conversion-doesnt-work-on-orin-nano/380004)

<div class="topic-metadata">

**Author:** [@adit\_bhrgv](https://forums.developer.nvidia.com/u/adit_bhrgv)\
**Replies:** 6\
**Last updated:** [August 28, 2026, 6:17am UTC](https://forums.developer.nvidia.com/t/onnx-to-trt-conversion-doesnt-work-on-orin-nano/380004 "2026-08-28T06:17:49Z")

</div>

Hi All, We are trying to convert an EfficientDet-D4 ONNX model to a TensorRT engine on Jetson Orin Nano 8GB custom board using JetPack 6.2, CUDA 12.6, and TensorRT 10.3.0. Model Details TF → ONNX using tf2onnx ONNX 1.…

---

## [AGX orin 32GB JP7.2 CUDA and tensorRT problems](https://forums.developer.nvidia.com/t/agx-orin-32gb-jp7-2-cuda-and-tensorrt-problems/378862)

<div class="topic-metadata">

**Author:** [@smileandcry2023](https://forums.developer.nvidia.com/u/smileandcry2023)\
**Replies:** 5\
**Last updated:** [August 28, 2026, 3:12am UTC](https://forums.developer.nvidia.com/t/agx-orin-32gb-jp7-2-cuda-and-tensorrt-problems/378862 "2026-08-28T03:12:52Z")

</div>

Hi HW: AGX Orin 32GB JP: 7.2 I have several problems about AGX Orin 32GB with JetPack7.2 for super mode. 1, Can I install the cuda-toolkit-13.2.2 instead of cuda-toolkit-13.2.1 ? We found a bug would decrease the …

---

## [Are TensorRT engines portable across identical Jetson Orin NX devices, or must they be built per device?](https://forums.developer.nvidia.com/t/are-tensorrt-engines-portable-across-identical-jetson-orin-nx-devices-or-must-they-be-built-per-device/381300)

<div class="topic-metadata">

**Author:** [@maimoon](https://forums.developer.nvidia.com/u/maimoon)\
**Replies:** 2\
**Last updated:** [August 27, 2026, 5:21am UTC](https://forums.developer.nvidia.com/t/are-tensorrt-engines-portable-across-identical-jetson-orin-nx-devices-or-must-they-be-built-per-device/381300 "2026-08-27T05:21:56Z")

</div>

Setup: Jetson Orin NX 16GB, JetPack 6.2 target, TensorRT as shipped with that JetPack. Pre-deployment design question. We plan to update detection models over the air to unattended remote sites with limited bandwidth an…

---

## [CUDA sporadically Fails to Deserialize Engine File](https://forums.developer.nvidia.com/t/cuda-sporadically-fails-to-deserialize-engine-file/380546)

<div class="topic-metadata">

**Author:** [@alexander.s.dodd](https://forums.developer.nvidia.com/u/alexander.s.dodd)\
**Replies:** 0\
**Last updated:** [August 18, 2026, 10:28pm UTC](https://forums.developer.nvidia.com/t/cuda-sporadically-fails-to-deserialize-engine-file/380546 "2026-08-18T22:28:55Z")

</div>

We are running CUDA Version 13.1 with driver version 590.48.01 on a AWS G5.24xlarge with TensorRT version 10.14.1. We built the engine on the same box and it will load some times but not others. The error we get when we…

---

## [算法导致 thor 频现oc3报警](https://forums.developer.nvidia.com/t/thor-oc3/379103)

<div class="topic-metadata">

**Author:** [@liu.jialu](https://forums.developer.nvidia.com/u/liu.jialu)\
**Replies:** 10\
**Last updated:** [August 13, 2026, 7:25am UTC](https://forums.developer.nvidia.com/t/thor-oc3/379103 "2026-08-13T07:25:32Z")

</div>

硬件平台：thor 官方开发套件 软件平台：jetpack 7.1 问题复现软件包： thor\_depthnet3\_oc3\_supplier\_handoff\_20260801.zip (14.3 MB) NVIDIA Thor BEVFusion DepthNet OC3 问题复现包 1. 问题摘要 平台：NVIDIA Thor，Jetson Linux R38.4，Ubuntu 24.04，CUDA 13.0，TensorR…

---

## [TRT Engine plan file across devices warning after reboot on Jetson Orin Nano](https://forums.developer.nvidia.com/t/trt-engine-plan-file-across-devices-warning-after-reboot-on-jetson-orin-nano/376848)

<div class="topic-metadata">

**Author:** [@allenxia](https://forums.developer.nvidia.com/u/allenxia)\
**Replies:** 2\
**Last updated:** [August 11, 2026, 8:20am UTC](https://forums.developer.nvidia.com/t/trt-engine-plan-file-across-devices-warning-after-reboot-on-jetson-orin-nano/376848 "2026-08-11T08:20:01Z")

</div>

I am experiencing an issue with TensorRT engine compatibility on both Jetson Orin Nano Super and non-Super boards throwing a Using an engine plan file across different models of devices is not supported and is likely to …

---

## [Mission Control — Autonomous AI Gaming Assistant & Telemetry Control powered by NVIDIA NIM & TensorRT](https://forums.developer.nvidia.com/t/mission-control-autonomous-ai-gaming-assistant-telemetry-control-powered-by-nvidia-nim-tensorrt/379635)

<div class="topic-metadata">

**Author:** [@missioncontrolgg](https://forums.developer.nvidia.com/u/missioncontrolgg)\
**Replies:** 0\
**Last updated:** [August 9, 2026, 1:00pm UTC](https://forums.developer.nvidia.com/t/mission-control-autonomous-ai-gaming-assistant-telemetry-control-powered-by-nvidia-nim-tensorrt/379635 "2026-08-09T13:00:54Z")

</div>

Hi NVIDIA Developer Community & DevRel Team, I’m Arnab, creator of Mission Control — an autonomous AI gaming assistant, HUD overlay, and hardware telemetry control dashboard optimized specifically for NVIDIA GeForce RTX…

---

## [Techniques to Imporve TensorRT Model Inference Speed](https://forums.developer.nvidia.com/t/techniques-to-imporve-tensorrt-model-inference-speed/331917)

<div class="topic-metadata">

**Author:** [@arumugamdeepakkumar](https://forums.developer.nvidia.com/u/arumugamdeepakkumar)\
**Replies:** 3\
**Last updated:** [August 5, 2026, 1:24am UTC](https://forums.developer.nvidia.com/t/techniques-to-imporve-tensorrt-model-inference-speed/331917 "2026-08-05T01:24:55Z")

</div>

Description Ways to improve Model Inference Speed We aim to enhance the inference speed of our object detection model without compromising accuracy. Current Setup: Model Task: Object Detection (3 Classes) Hardware: N…

---

## [Compute bottleneck evaluating multiple trajectories through a 7B VLM (VLA Architecture Design)](https://forums.developer.nvidia.com/t/compute-bottleneck-evaluating-multiple-trajectories-through-a-7b-vlm-vla-architecture-design/374001)

<div class="topic-metadata">

**Author:** [@6farsi](https://forums.developer.nvidia.com/u/6farsi)\
**Replies:** 4\
**Last updated:** [August 4, 2026, 10:59am UTC](https://forums.developer.nvidia.com/t/compute-bottleneck-evaluating-multiple-trajectories-through-a-7b-vlm-vla-architecture-design/374001 "2026-08-04T10:59:29Z")

</div>

Hi everyone, I am an undergraduate researcher building a new Reasoning-Consistent VLA (Vision-Language-Action) framework for autonomous driving. To give some background, I recently completed and open-sourced an autonom…

[Next page](https://forums.developer.nvidia.com/tag/tensorrt/12.md?match_all_tags=true&page=1&tags%5B%5D=tensorrt)
