From 7788a6b84994dbb16ba607ea87c8ef35d1306a23 Mon Sep 17 00:00:00 2001 From: Hugo Larcher Date: Wed, 4 Dec 2024 10:49:41 +0100 Subject: [PATCH] doc: Formatting --- docs/source/multi_backend_support.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/source/multi_backend_support.md b/docs/source/multi_backend_support.md index d8c8ee2e9d9..5899e4b77d4 100644 --- a/docs/source/multi_backend_support.md +++ b/docs/source/multi_backend_support.md @@ -6,8 +6,8 @@ whether you prioritize performance, ease of use, or compatibility with specific TGI remains consistent across backends, allowing you to switch between them seamlessly. **Supported backends:** -* TGI CUDA backend: This high-performance backend is optimized for NVIDIA GPUs and serves as the default option +* **TGI CUDA backend**: This high-performance backend is optimized for NVIDIA GPUs and serves as the default option within TGI. Developed in-house, it boasts numerous optimizations and is used in production by various projects, including those by Hugging Face. -* [TGI TRTLLM backend](./backends/trtllm): This backend leverages NVIDIA's TensorRT library to accelerate LLM inference. +* **[TGI TRTLLM backend](./backends/trtllm)**: This backend leverages NVIDIA's TensorRT library to accelerate LLM inference. It utilizes specialized optimizations and custom kernels for enhanced performance. However, it requires a model-specific compilation step for each GPU architecture. \ No newline at end of file