TensorRT-LLM Inference

What it does

When your agent uses it

• Deploying a model for production on NVIDIA GPUs with low latency
• Serving a quantized FP8 or INT4 model with in-flight batching
• Splitting one large model across several GPUs or nodes
• Deciding between TensorRT-LLM, vLLM and llama.cpp for a deployment

Example prompts

• “Start trtllm-serve for Llama 3 8B with tensor parallelism across four GPUs.”
• “Load an FP8 quantized model with TensorRT-LLM and run a batch of prompts.”
• “Write a Docker command to run the TensorRT-LLM image with GPU access.”
• “Should we use TensorRT-LLM or vLLM for our single-node deployment?”

Requirements

• NVIDIA GPUs
• Docker, or a TensorRT-LLM installation with Python

What it can do on your machine

Read from SKILL.md and the folder at commit 773a529. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Shell commands in SKILL.md call:

  • docker
  • pip
  • curl

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  Links to these hosts (documentation or services it may open):

  • nvidia.github.io
  • github.com
  • huggingface.co

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

TensorRT-LLM Inference loads about 1.3k tokens when it runs, and up to ~6.8k if it reads all its reference files. Until then it costs about 77 tokens; SKILL.md has 284 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~77

When it runs · the whole SKILL.md, loaded when a task matches

~1.3k

With references · SKILL.md plus every file in references/, read only if the agent opens them

~6.8k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from Orchestra-Research/AI-Research-SKILLs at commit 773a529, republished under its MIT licence (© Orchestra-Research). 284 words, ~1,259 tokens.

name

tensorrt-llm

description

Optimizes LLM inference with NVIDIA TensorRT for maximum throughput and lowest latency. Use for production deployment on NVIDIA GPUs (A100/H100), when you need 10-100x faster inference than PyTorch, or for serving models with quantization (FP8/INT4), in-flight batching, and multi-GPU scaling.

version

1.0.0

author

Orchestra Research

license

MIT

tags

Inference Serving, TensorRT-LLM, NVIDIA, Inference Optimization, High Throughput, Low Latency, Production, FP8, INT4, In-Flight Batching, Multi-GPU

dependencies

tensorrt-llm, torch

# Docker (recommended)
docker pull nvidia/tensorrt_llm:latest

# pip install
pip install tensorrt_llm==1.2.0rc3

# Requires CUDA 13.0.0, TensorRT 10.13.2, Python 3.10-3.12

from tensorrt_llm import LLM, SamplingParams

# Initialize model
llm = LLM(model="meta-llama/Meta-Llama-3-8B")

# Configure sampling
sampling_params = SamplingParams(
    max_tokens=100,
    temperature=0.7,
    top_p=0.9
)

# Generate
prompts = ["Explain quantum computing"]
outputs = llm.generate(prompts, sampling_params)

for output in outputs:
    print(output.text)

# Start server (automatic model download and compilation)
trtllm-serve meta-llama/Meta-Llama-3-8B \
    --tp_size 4 \              # Tensor parallelism (4 GPUs)
    --max_batch_size 256 \
    --max_num_tokens 4096

# Client request
curl -X POST http://localhost:8000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "meta-llama/Meta-Llama-3-8B",
    "messages": [{"role": "user", "content": "Hello!"}],
    "temperature": 0.7,
    "max_tokens": 100
  }'

from tensorrt_llm import LLM

# Load FP8 quantized model (2× faster, 50% memory)
llm = LLM(
    model="meta-llama/Meta-Llama-3-70B",
    dtype="fp8",
    max_num_tokens=8192
)

# Inference same as before
outputs = llm.generate(["Summarize this article..."])

# Tensor parallelism across 8 GPUs
llm = LLM(
    model="meta-llama/Meta-Llama-3-405B",
    tensor_parallel_size=8,
    dtype="fp8"
)

# Process 100 prompts efficiently
prompts = [f"Question {i}: ..." for i in range(100)]

outputs = llm.generate(
    prompts,
    sampling_params=SamplingParams(max_tokens=200)
)

# Automatic in-flight batching for maximum throughput

© Orchestra-Research, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 3 other files (references) in 12-inference-serving/tensorrt-llm of Orchestra-Research/AI-Research-SKILLs.

• SKILL.md
• references/multi-gpu.md
• references/optimization.md
• references/serving.md

Open the folder on GitHubat commit 773a529

Used in 4 other repositories

We found 4 copies of this SKILL.md (exact, near-identical or edited) in other folders, from 4 other GitHub owners. This page covers the copy in Orchestra-Research/AI-Research-SKILLs, which our catalogue first saw on October 7, 2026.

• Tommy-yw/RunbookHermesoptional-skills/mlops/tensorrt-llm
• davila7/claude-code-templatescli-tool/components/skills/ai-research/inference-serving-tensorrt-llm
• johnson7788/MultiUserClawhermes-agent/optional-skills/mlops/tensorrt-llm
• OpenLAIR/dr-clawskills/inference-serving/tensorrt-llm