NVIDIA/TensorRT-Model-Connect

View on GitHub

From PyTorch model to end-to-end TensorRT inference experience in two commands—AI-native, cross-platform, and built for the best possible user experience.

Stars: 254
Forks: 58
Open beginner issues: 2
Indexed issues: 47
Avg merge: 1d 8h
Merged PRs (30d): 223

Dominant language: Python
License: Apache-2.0
Last GitHub push: Sep 19, 2026
Latest indexed: Sep 19, 2026
Contributing guide: Contributing guide
Code of conduct: No code of conduct
Beginner labels: good first issue

Beginner friendly Everything

Open All

47 issues indexed so far

Feature: Add Qwen3-VL-8B-Instruct to the qwen_vl family Open

Difficulty 4/5 3-5 days Newbie friendliness 68/100

NVIDIA/TensorRT-Model-Connect#1388 · 1 comment · Opened Sep 20, 2026
Feature: Add point-cloud semantic segmentation task and PointNet model family Open

enhancement

Difficulty 5/5 Over a week Newbie friendliness 35/100

NVIDIA/TensorRT-Model-Connect#1328 · 1 comment · Opened Sep 17, 2026
Feature: Enable Telemetry for TRTMC Open

New Feature Request Roadmap

Difficulty 5/5 Over a week Newbie friendliness 35/100

NVIDIA/TensorRT-Model-Connect#1325 · 3 comments · Opened Sep 17, 2026
Feature: Add Fun-CosyVoice3-0.5B-2512 native TensorRT support Open

Difficulty 5/5 Over a week Newbie friendliness 30/100

NVIDIA/TensorRT-Model-Connect#1263 · 2 comments · Opened Sep 11, 2026
Feature: preserve multichannel audio in public results and WAV output Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

NVIDIA/TensorRT-Model-Connect#1254 · Opened Sep 10, 2026
Question: Interest in HeartMuLa 3B music-generation support? Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

NVIDIA/TensorRT-Model-Connect#1227 · 1 comment · Opened Sep 9, 2026
Cross-family runtime safety and long-context issues found during #1158 review Open

Difficulty 5/5 Over a week Newbie friendliness 35/100

NVIDIA/TensorRT-Model-Connect#1177 · 1 comment · Opened Sep 5, 2026
Two defects every model family inherited: an unguarded device query and GPU tests that pass without running Open

Difficulty 5/5 Over a week Newbie friendliness 35/100

NVIDIA/TensorRT-Model-Connect#1176 · Opened Sep 5, 2026
Docs: ITrtModule::forward() output-buffer lifetime and dtype contract are undocumented Open

Difficulty 2/5 1-3 hours Newbie friendliness 82/100

NVIDIA/TensorRT-Model-Connect#1126 · 1 comment · Opened Sep 2, 2026
Docs: repo agent skills are unreachable from every documented entry point Open

Difficulty 2/5 1-3 hours Newbie friendliness 85/100

NVIDIA/TensorRT-Model-Connect#1125 · 1 comment · Opened Sep 2, 2026
[Enterprise] Add end-to-end ESMFold2 support Open

enhancement Enterprise

Difficulty 5/5 Over a week Newbie friendliness 28/100

NVIDIA/TensorRT-Model-Connect#1113 · Opened Aug 31, 2026
[Enterprise] Complete MiniMax H3 production qualification Open

Enterprise

Difficulty 5/5 Over a week Newbie friendliness 28/100

NVIDIA/TensorRT-Model-Connect#1109 · 1 comment · Opened Aug 31, 2026
[Community] Prepare TensorRT Model Connect for GTC Open

Community

Difficulty 5/5 Over a week Newbie friendliness 30/100

NVIDIA/TensorRT-Model-Connect#1108 · Opened Aug 31, 2026
[Demo] Create the canonical three-step Model Connect demo Open

Community good first issue

Difficulty 5/5 Over a week Newbie friendliness 48/100

NVIDIA/TensorRT-Model-Connect#1107 · 3 comments · Opened Aug 31, 2026
[Community] Curate and protect a queue of good first issues Open

Community

Difficulty 5/5 Over a week Newbie friendliness 35/100

NVIDIA/TensorRT-Model-Connect#1106 · Opened Aug 31, 2026
[Community] Launch recurring Model Connect livestreams Open

Community

Difficulty 5/5 Over a week Newbie friendliness 35/100

NVIDIA/TensorRT-Model-Connect#1105 · Opened Aug 31, 2026
[Benchmark] Integrate the ALE benchmark Open

Platform

Difficulty 5/5 Over a week Newbie friendliness 35/100

NVIDIA/TensorRT-Model-Connect#1104 · 1 comment · Opened Aug 31, 2026
[Community] Add a Hugging Face badge for verified Model Connect support Open

Community

Difficulty 5/5 Over a week Newbie friendliness 25/100

NVIDIA/TensorRT-Model-Connect#1103 · Opened Aug 31, 2026
[Community] Establish QA staffing and ownership Open

Community

Difficulty 5/5 Over a week Newbie friendliness 28/100

NVIDIA/TensorRT-Model-Connect#1102 · Opened Aug 31, 2026
[Community] Provision GPU capacity for Community CI Open

Community

Difficulty 5/5 Over a week Newbie friendliness 25/100

NVIDIA/TensorRT-Model-Connect#1101 · Opened Aug 31, 2026
[Docs] Add TensorRT Model Connect to the TensorRT documentation Open

Platform

Difficulty 4/5 3-5 days Newbie friendliness 68/100

NVIDIA/TensorRT-Model-Connect#1100 · 1 comment · Opened Aug 31, 2026
[Platform] Add qualified runtime dispatch to Edge-LLM Open

Platform

Difficulty 5/5 Over a week Newbie friendliness 30/100

NVIDIA/TensorRT-Model-Connect#1099 · Opened Aug 31, 2026
[Platform] Enhance TVM-FFI Bridge and open up kernel slots for flashinfer Open

Platform

Difficulty 5/5 Over a week Newbie friendliness 25/100

NVIDIA/TensorRT-Model-Connect#1098 · 1 comment · Opened Aug 31, 2026
[Platform] Identify and selectively deduplicate shared code hot spots Open

Platform

Difficulty 5/5 Over a week Newbie friendliness 35/100

NVIDIA/TensorRT-Model-Connect#1097 · Opened Aug 31, 2026
Feature: Add MiniMaxAI/MiniMax-Music3 support Open

AI enhancement model-onboarding

NVIDIA/TensorRT-Model-Connect#1092 · 8 comments · 1 assignee · Opened Aug 29, 2026
[RFC] Selective debug windows in dual-profile builders - extend the sana_wm/gemma pattern to generic families Open

Difficulty 5/5 Over a week Newbie friendliness 38/100

NVIDIA/TensorRT-Model-Connect#1088 · 2 comments · Opened Aug 29, 2026
Bug: checkpoint validation disappears under optimized Python Open

Difficulty 5/5 Over a week Newbie friendliness 45/100

NVIDIA/TensorRT-Model-Connect#1052 · 3 comments · Opened Aug 26, 2026
Feature: Add ornith-ai/Ornith-1.5-9B checkpoint support Open

AI enhancement model-onboarding

NVIDIA/TensorRT-Model-Connect#1050 · 2 comments · 1 assignee · Opened Aug 26, 2026
Question: What is the runtime of Fast-FoundationStereo on a 5090? Open

question

Difficulty 3/5 1-2 days Newbie friendliness 35/100

NVIDIA/TensorRT-Model-Connect#1045 · 1 comment · Opened Aug 26, 2026
Feature: Add Wan 3.0 video model support Open

AI enhancement good first issue

NVIDIA/TensorRT-Model-Connect#1038 · 3 comments · 1 assignee · Opened Aug 26, 2026
Feature: Introduce Node.js Language Bindings for TRT-Model-Connect (Prototype) Open

enhancement

Difficulty 5/5 Over a week Newbie friendliness 30/100

NVIDIA/TensorRT-Model-Connect#1030 · 1 comment · Opened Aug 25, 2026
Feature: support biology models - Boltz-2, OpenFold3, ESMFold2 Open

enhancement Enterprise

Difficulty 5/5 Over a week Newbie friendliness 25/100

NVIDIA/TensorRT-Model-Connect#998 · 3 comments · Opened Aug 21, 2026
Docs: Clarify the quantization → inference → serving workflow in TRT Model Connect, with clear end-to-end user documentation. Open

Difficulty 4/5 3-5 days Newbie friendliness 45/100

NVIDIA/TensorRT-Model-Connect#988 · 2 comments · Opened Aug 21, 2026
Feature: Visibility tools for perf optimization and debugging Open

enhancement New Feature Request

Difficulty 5/5 Over a week Newbie friendliness 28/100

NVIDIA/TensorRT-Model-Connect#987 · Opened Aug 21, 2026
Feature: Enable a self-service import path for customer-proprietary models, allowing customers to import their models without sharing model architecture or weights with NVIDIA. Open

New Feature Request

Difficulty 5/5 Over a week Newbie friendliness 25/100

NVIDIA/TensorRT-Model-Connect#986 · Opened Aug 21, 2026
[Refactor] Render E2E reproduction commands with shlex.join Open

enhancement good first issue

Difficulty 4/5 3-5 days Newbie friendliness 56/100

NVIDIA/TensorRT-Model-Connect#976 · 3 comments · Opened Aug 21, 2026
Question: Qwen3 native KV-cache decode engine fails to build on SM120/SM121 (DGX Spark / GB10) Open

bug question

Difficulty 4/5 3-5 days Newbie friendliness 42/100

NVIDIA/TensorRT-Model-Connect#955 · 10 comments · Opened Aug 20, 2026
Feature: Make documentation LLM friendly Open

enhancement New Feature Request

Difficulty 5/5 Over a week Newbie friendliness 35/100

NVIDIA/TensorRT-Model-Connect#954 · 1 comment · Opened Aug 20, 2026
Feature: LFM2-MoE and LFM2-VL follow-on families (scoping) Open

enhancement

Difficulty 5/5 Over a week Newbie friendliness 35/100

NVIDIA/TensorRT-Model-Connect#943 · 4 comments · Opened Aug 20, 2026
[Performance] nemotron-hindi-4b latency is 17.0% above reference Open

AI bug

Difficulty 4/5 3-5 days Newbie friendliness 42/100

NVIDIA/TensorRT-Model-Connect#905 · Opened Aug 17, 2026
[Performance] glm-4-9b latency is 25.5% above reference Open

AI bug nvbug:trtmc

Difficulty 4/5 3-5 days Newbie friendliness 42/100

NVIDIA/TensorRT-Model-Connect#904 · Opened Aug 17, 2026
[Performance] riva-translate-4b latency is 35.3% above reference Open

AI bug

Difficulty 5/5 Over a week Newbie friendliness 38/100

NVIDIA/TensorRT-Model-Connect#903 · Opened Aug 17, 2026
[Performance] qwen3-4b-instruct-2507 latency is 39.7% above reference Open

AI bug

Difficulty 4/5 3-5 days Newbie friendliness 48/100

NVIDIA/TensorRT-Model-Connect#901 · Opened Aug 17, 2026
[Performance] mistral-7b latency is 65.5% above the FP16 reference Open

AI bug

Difficulty 4/5 3-5 days Newbie friendliness 48/100

NVIDIA/TensorRT-Model-Connect#900 · Opened Aug 17, 2026
[Performance] phi-moe is 3.84x slower than its FP16 reference Open

AI bug

Difficulty 4/5 3-5 days Newbie friendliness 45/100

NVIDIA/TensorRT-Model-Connect#895 · Opened Aug 17, 2026
Bug: TensorRT 11.2 bundles can hit Myelin CUDA illegal memory access after successful generation Open

AI bug

Difficulty 5/5 Over a week Newbie friendliness 32/100

NVIDIA/TensorRT-Model-Connect#428 · Opened Jul 9, 2026
CI: provide a non-publishing PR lane for clean-wheel model smoke Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 45/100

NVIDIA/TensorRT-Model-Connect#406 · Opened Jul 7, 2026

NVIDIA/TensorRT-Model-Connect

Get new issues in your inbox