Skip to content

Releases: openvinotoolkit/openvino_testdrive

v24.6.0

19 Dec 17:02
@AlexanderBarabanov AlexanderBarabanov
v24.6.0.1
6247b4d
Compare
Choose a tag to compare
v24.6.0 Pre-release
Pre-release

Summary of major features and improvements

  • Image generation pipeline: OpenVINO™ Test Drive now supports OpenVINO™ GenAI API-powered image generation tool with an LLM-like UI to enable users to test Stable Diffusion models, copy generated images, and monitor performance metrics on various devices.
  • Video transcription generation pipeline: audio-to-text generation pipeline using OpenVINO™ GenAI to enable users to select videos, generate live subtitles, full transcriptions, and measure performance metrics on specified target devices.
  • UX Improvements:
    • Model manifest was added to enrich model card contents with details such as model description, model size, task type, and optimization.
    • UX and navigation menu redesign, feedback button was added, light and dark modes are supported now.
    • Model export feature was implemented to allow user to save model to custom directory for future usage.
  • Upgrade to OpenVINO 2024.6.

Download this release

Download windows installer.

Assets 2
Loading

v24.4.0

06 Nov 08:41
@RHeckerIntel RHeckerIntel
v24.4.0
24e4015
Compare
Choose a tag to compare
v24.4.0 Pre-release
Pre-release

🚀 [v24.4.0]

Summary of major features

  • Download LLM Models: Access models from the Huggingface OpenVINO LLM collection.
  • Text Generation: Utilize LLM models for generating text.
  • Chat Templates: Implement LLM with pre-defined chat templates.
  • Performance Metrics: Monitor LLM model performance, including load time, time to first token, and time per output token.
  • Configuration Options: Adjust temperature and Top P settings to control the variability and randomness of generated responses.
  • Inference for Computer Vision Models: Run inference on all computer vision models trained using Intel Geti.
  • Task-Chain Inference: Execute inference for all task-chains created by Intel Geti.
  • Model Testing: Test computer vision models using a sample image or an image from the local disk.
  • Batch Inference: Perform batch inference for computer vision by selecting source and destination folders.
  • Processor Selection: Choose processor types (CPU, GPU, NPU) for model execution.
  • Windows Installer: Install via MSIX for a seamless setup.

Known issues

  • An auto plugin might trigger issues in a large language model (LLM).

Download this release

Download windows installer of V24.4.0 beta release of OpenVINO Test Drive

Assets 2
Loading