jalr/tabbyAPI-ollama
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
No description
861 commits 1 branch 0 tags 1.4 MiB
Find a file
kingbri bd16681825 Start: Mark cuda 11.8 as unsupported 
Temporary until existing cuda 11.8 scripts can be migrated to cuda 12.

Signed-off-by: kingbri <8082010+bdashore3@users.noreply.github.com>
2025-01-12 21:50:41 -05:00
.github fix issues with optional dependencies (#204) 2024-09-19 22:24:55 -04:00
backends Model: Cleanup generation length and page checks 2024-12-26 23:13:08 -05:00
colab Model: Remove override_base_seq_len 2024-10-30 10:03:08 +08:00
common Dependencies: Remove outlines from optional check 2024-12-18 11:56:40 -05:00
docker Docker: Add extras to dockerfile 2024-11-15 18:16:48 -05:00
endpoints Embeddings: Fix base64 return 2025-01-01 16:15:12 -05:00
loras Implement lora support (#24) 2023-12-08 23:38:08 -05:00
models Tree: Update documentation and configs 2023-11-16 02:30:33 -05:00
sampler_overrides Sampling: Add XTC support 2024-09-24 18:10:52 -04:00
templates Templates: Alter chatml_with_headers to fit huggingface spec 2024-12-30 14:00:44 -05:00
tests Tree: Format 2024-03-13 00:02:55 -04:00
update_scripts Start: Make linux scripts executable 2024-08-03 15:19:31 -04:00
.dockerignore debloat docker build 2024-09-08 00:02:00 +01:00
.gitignore Update .gitignore 2024-09-16 22:48:13 -04:00
api_tokens_sample.yml Improve docker deployment configuration (#163) 2024-08-18 15:19:18 -04:00
config_sample.yml Merge branch 'main' into vision 2024-11-19 21:19:21 -05:00
formatting.bat feat: workflows for formatting/linting (#35) 2023-12-22 16:20:35 +00:00
formatting.sh feat: workflows for formatting/linting (#35) 2023-12-22 16:20:35 +00:00
LICENSE Create LICENSE 2023-11-16 17:43:23 -05:00
main.py Model: Fix inline loading and draft key (#225) 2024-10-24 23:35:05 -04:00
pyproject.toml Bump formatron to 0.4.11 2025-01-08 00:48:25 +01:00
README.md Update README 2024-08-17 00:35:42 -04:00
start.bat Start: Fix startup with new argparser 2024-09-21 14:36:21 -04:00
start.py Start: Mark cuda 11.8 as unsupported 2025-01-12 21:50:41 -05:00
start.sh Start: Fix startup with new argparser 2024-09-21 14:36:21 -04:00

README.md

TabbyAPI

Python 3.10, 3.11, and 3.12 License: AGPL v3 Discord Server

Developer facing API documentation

Support on Ko-Fi

Important

In addition to the README, please read the Wiki page for information about getting started!

Note

Need help? Join the Discord Server and get the Tabby role. Please be nice when asking questions.

A FastAPI based application that allows for generating text using an LLM (large language model) using the Exllamav2 backend

TabbyAPI is also the official API backend server for ExllamaV2.

Disclaimer

This project is marked as rolling release. There may be bugs and changes down the line. Please be aware that you might need to reinstall dependencies if needed.

TabbyAPI is a hobby project made for a small amount of users. It is not meant to run on production servers. For that, please look at other solutions that support those workloads.

Getting Started

Important

This README does not have instructions for setting up. Please read the Wiki.

Read the Wiki for more information. It contains user-facing documentation for installation, configuration, sampling, API usage, and so much more.

Features

  • OpenAI compatible API
  • Loading/unloading models
  • HuggingFace model downloading
  • Embedding model support
  • JSON schema + Regex + EBNF support
  • AI Horde support
  • Speculative decoding via draft models
  • Multi-lora with independent scaling (ex. a weight of 0.9)
  • Inbuilt proxy to override client request parameters/samplers
  • Flexible Jinja2 template engine for chat completions that conforms to HuggingFace
  • Concurrent inference with asyncio
  • Utilizes modern python paradigms
  • Continuous batching engine using paged attention
  • Fast classifer-free guidance
  • OAI style tool/function calling

And much more. If something is missing here, PR it in!

Supported Model Types

TabbyAPI uses Exllamav2 as a powerful and fast backend for model inference, loading, etc. Therefore, the following types of models are supported:

  • Exl2 (Highly recommended)

  • GPTQ

  • FP16 (using Exllamav2's loader)

In addition, TabbyAPI supports parallel batching using paged attention for Nvidia Ampere GPUs and higher.

Contributing

Use the template when creating issues or pull requests, otherwise the developers may not look at your post.

If you have issues with the project:

  • Describe the issue in detail

  • If you have a feature request, please indicate it as such.

If you have a Pull Request

  • Describe the pull request in detail, what, and why you are changing something

Acknowldgements

TabbyAPI would not exist without the work of other contributors and FOSS projects:

Developers and Permissions

Creators/Developers: