mirror of
https://github.com/TabbyML/tabby
synced 2024-11-22 00:08:06 +00:00
docs: create blog post for vulkan support (#2021)
* docs: create blog post for vulkan support * Update website/blog/2024-05-01-vulkan-support/index.md --------- Co-authored-by: Meng Zhang <meng@tabbyml.com>
This commit is contained in:
parent
7bac39e945
commit
8d8c45c393
BIN
website/blog/2024-05-01-vulkan-support/completion.png
(Stored with Git LFS)
vendored
Normal file
BIN
website/blog/2024-05-01-vulkan-support/completion.png
(Stored with Git LFS)
vendored
Normal file
Binary file not shown.
61
website/blog/2024-05-01-vulkan-support/index.md
vendored
Normal file
61
website/blog/2024-05-01-vulkan-support/index.md
vendored
Normal file
@ -0,0 +1,61 @@
|
||||
---
|
||||
title: 'Vulkan Support: LLMs for Everyone'
|
||||
authors: [boxbeam]
|
||||
tags: [deployment]
|
||||
---
|
||||
|
||||
It has long been the case that machine learning models are run on the GPU to improve their performance.
|
||||
The GPU is far more effective at the kinds of computations needed for AI than the CPU, and so GPU compute libraries
|
||||
such as Cuda and ROCm are typically used.
|
||||
|
||||
However, requiring the support of these libraries can restrict which graphics cards are compatible, leaving many
|
||||
with older or less popular cards unable to run LLMs efficiently.
|
||||
|
||||
Tabby is happy to announce that we now support Vulkan, a graphics library created primarily for games. Its original purpose
|
||||
means that it is designed to work on a very broad range of cards, and leveraging it to host LLMs means that we can now
|
||||
offer GPU acceleration to people whose cards are not supported by Cuda and ROCm.
|
||||
|
||||
Vulkan works on basically any GPU, so if you have previously been forced to host local models on your CPU, now is the time
|
||||
to see what Tabby with Vulkan can do for you!
|
||||
|
||||
## Vulkan Installation
|
||||
|
||||
To begin, first make sure that you have Vulkan installed.
|
||||
|
||||
For Windows users, Vulkan may be natively supported. Otherwise, the Vulkan SDK can be downloaded at https://vulkan.lunarg.com/sdk/home#windows.
|
||||
|
||||
For Linux users, Vulkan can be installed through your package manager:
|
||||
- Arch Linux: vulkan-icd-loader (universal), and also install vulkan-radeon (for AMD) or vulkan-nouveau (for Nvidia)
|
||||
- Debian Linux: libvulkan1
|
||||
|
||||
![Vulkan installed on Arch Linux](./vulkan-installed-on-arch.png)
|
||||
|
||||
## Tabby Installation
|
||||
|
||||
To start using Tabby with Vulkan, first download one of the pre-built Vulkan binaries for your platform:
|
||||
- Linux: https://github.com/TabbyML/tabby/releases/download/v0.10.0/tabby_x86_64-manylinux2014-vulkan
|
||||
- Windows: https://github.com/TabbyML/tabby/releases/download/v0.10.0/tabby_x86_64-windows-msvc-vulkan.exe
|
||||
|
||||
## Running
|
||||
|
||||
Once you've installed the appropriate binary, you can simply run it from the command line:
|
||||
|
||||
For Windows, open a command prompt and navigate to the download folder, then run:
|
||||
|
||||
```
|
||||
tabby_x84_64-windows-msvc-vulkan serve --model StarCoder-1B --device vulkan
|
||||
```
|
||||
|
||||
For Linux:
|
||||
|
||||
```
|
||||
./tabby_x64_64-manylinux2014-vulkan serve --model StarCoder-1B --device vulkan
|
||||
```
|
||||
|
||||
When it starts, you should see a printout indicating that Vulkan has found your card and is working properly:
|
||||
|
||||
![Vulkan running on Linux](./vulkan-running.png)
|
||||
|
||||
Now enjoy your speedy completions!
|
||||
|
||||
![Completion example](./completion.png)
|
BIN
website/blog/2024-05-01-vulkan-support/vulkan-installed-on-arch.png
(Stored with Git LFS)
vendored
Normal file
BIN
website/blog/2024-05-01-vulkan-support/vulkan-installed-on-arch.png
(Stored with Git LFS)
vendored
Normal file
Binary file not shown.
BIN
website/blog/2024-05-01-vulkan-support/vulkan-running.png
(Stored with Git LFS)
vendored
Normal file
BIN
website/blog/2024-05-01-vulkan-support/vulkan-running.png
(Stored with Git LFS)
vendored
Normal file
Binary file not shown.
Loading…
Reference in New Issue
Block a user