Skip to content

Releases: withmartian/llm-adapters

v8.0.1

04 Mar 05:35
6c3890c
Compare
Choose a tag to compare

What's Changed

Full Changelog: v7.9.0...v8.0.1

Moved to new format, now we use OpenAI SDK format

# OpenAI

from openai import OpenAI

client = OpenAI()
completion = client.chat.completions.create(
    model="gpt-4o",
    messages=[
        {"role": "developer", "content": "You are a helpful assistant."},
    ]
)

# New Adapters

from llm_adapters import OpenAI

client = OpenAI()
completion = client.chat.completions.create(
    model="gpt-4o",
    messages=[
        {"role": "developer", "content": "You are a helpful assistant."},
    ]
)

v7.9.0

27 Feb 23:07
bef3be8
Compare
Choose a tag to compare

What's Changed

Full Changelog: v7.8.0...v7.9.0

v7.8.0

24 Feb 21:23
7946e78
Compare
Choose a tag to compare

What's Changed

New Contributors

Full Changelog: v7.4.0...v7.8.0

v7.4.0

07 Jan 12:34
21dd735
Compare
Choose a tag to compare

What's Changed

Full Changelog: v7.3.0...v7.4.0

v7.3.0

04 Jan 00:04
e2e7070
Compare
Choose a tag to compare

What's Changed

Full Changelog: v7.2.0...v7.3.0

v7.2.0

14 Dec 04:14
26f7444
Compare
Choose a tag to compare

Changes

🚀 Features

v7.1.0

11 Dec 19:17
0ec7384
Compare
Choose a tag to compare

What's Changed

Full Changelog: v7.0.0...v7.1.0

v7.0.0

09 Dec 20:47
b06cd9c
Compare
Choose a tag to compare

Changes

🚀 Features

v7.0.0a2

12 Nov 09:28
01a56a0
Compare
Choose a tag to compare

Changes

7.0.0a0

09 Nov 20:56
045bf44
Compare
Choose a tag to compare

What's Changed

Full Changelog: 6.11.0...7.0.0a0

  • Is backwards compatible
  • Move from circle ci to Github Actions
  • Adapters auto updating models tracking page
  • Stricter linting rules
  • Updated all dependencies to latest version, removed unused dependencies
  • Reduced bundle size
  • Moved to ruff, removed pylons
  • Started using all to expose public facing interface
  • Typed 100% internal functions (Some parts still use Any, and general dict)
  • Removed unused code
  • Added support for only system and assistant conversations
  • Removed function calling, in favor of tools
  • Together added vision support (Doesn’t support multiple images)
  • Cohere added streaming, function calling, json content and json output support
  • Cohere fix temperature scaling to OpenAI format
  • Fireworks add vision
  • And much more