v0.3.2 - Qwen2 VL and multilingual BLIP2

New Models

• New Model - BLIP2 Variant mBLIP and BLIP2 flan by @dnth in #67
• Qwen2 VL Integration by @dnth in #72

Bugfixes

• Pin dependencies version by @dnth in #70

Full Changelog: v0.3.1...v0.3.2

v0.3.1 - Patch model list only searching for packages in site-packages.

What's Changed

• Add TOC to README by @dnth in #65

Full Changelog: v0.3.0...v0.3.1

v0.3.0 - OpenAI Chat Completion API, YOLO Segmentation and Pose models

New Features

OpenAI Chat Completion compatible API.

A new endpoint has been added to support OpenAI's Chat Completion functionality

• Add OpenAI Chat Completion endpoint by @dnth in #61

Gradio Visualization

We've introduced a new feature that allows users to visualize results using Gradio4. This integration provides an intuitive and interactive way to display and explore model outputs, making it easier for users to understand and interpret their results.

• Visualize results in Gradio by @dnth in #60

New Models

YOLO Seg and Pose

Support for Ultralytics YOLO pose estimation and segmentation models has been added. This expansion of supported models allows users to work with more advanced computer vision tasks. Additionally, we've optimized the fuzzy matching system for unsupported models, improving the overall user experience when working with various model types.

• Suport ultralytics yolo pose and seg model. Optimized fuzzy matching for unsupported models. by @315386775 in #52

Bug fixes

• Open browser to fastapi docs page when running model_serve by @dnth in #54
• Fix soft import to exclude local folder by @dnth in #64

Breaking Changes

Result Class Abstraction

In this release, we've made a significant change by abstracting model outputs into a Result class. While this may require some adjustments to existing code, it provides a more consistent and robust way of handling model outputs across different types of models and tasks.

• Abstract model outputs into a Result class by @dnth in #58

Full Changelog: v0.2.1...v0.3.0

v0.2.1 - Ultralytics Classification Model and non blocking calls to Ray Serve

What's Changed

• Support ultralytics classification by @315386775 in #44
• Add non blocking calls to Ray Serve by @dnth in #51

New Contributors

• @315386775 made their first contribution in #44

Full Changelog: v0.2.0...v0.2.1

v0.2.0 - Ray Serve and FastAPI endpoint

What's Changed

• Ray Serve and FastAPI endpoint by @dnth in #50

Full Changelog: v0.1.3...v0.2.0

v0.1.3 - Phi 3.5 Vision

What's Changed

• Add Phi 3.5 Vision Model by @dnth in #43

Full Changelog: v0.1.2...v0.1.3

v0.1.2 - Florence 2 Model Series

What's Changed

• Add Florence 2 model series by Microsoft by @dnth in #42

Full Changelog: v0.1.1...v0.1.2

v0.1.1 - Patch Llama 3.2 Vision model

Full Changelog: v0.1.0...v0.1.1

v0.1.0 - Llama 3.2 Vision Model

What's Changed

• Add Llama 3.2 Vision Models by @dnth in #38
• Prefix the framework name before model_id by @dnth in #40

Full Changelog: v0.0.12...v0.1.0

v0.0.12 - Patch Ollama Python client Installation

pip install xinfer[all] installs ollama python client as well.