19 lines
544 B
Markdown
19 lines
544 B
Markdown
---
|
|
created: '2026-04-05T05:39:16+00:00'
|
|
id: c3851d78-9aa2-4a1f-b72b-4bd6c1221111
|
|
links: []
|
|
modified: '2026-04-05T05:39:16+00:00'
|
|
summary: 'Here''s a concise summary:
|
|
|
|
|
|
llama.cpp offers efficient CPU/GPU inference, vllm excels at tensor parallelism,
|
|
and ollama builds on llama.cpp, inheriting its advantages and limitations.'
|
|
tags: []
|
|
title: AI Information
|
|
---
|
|
|
|
llama.cpp is good for cpu/gpu inferencing
|
|
vllm is good for tensor parallelism
|
|
ollama sits on top of llama.cpp and shares most of its advantages and shortcomings
|
|
|
|
#ai #llama #vllm |