v0.1.84 - Supertonic Demo App

Supertonic TTS Support Added

This update brings Supertonic Text-to-Speech support to the TextDemo application
You can now generate high-quality speech directly inside the app with no extra setup required.

What’s New

• Supertonic TTS Integration: High-quality, expressive speech synthesis with fast inference.
• Real-time Parameter Controls: Adjust speed, steps, and style directly from the UI.
• Export to WAV: Save generated audio instantly.

Full Changelog: v0.1.81...v0.1.84

v0.1.81 - Supertonic Pipeline

What's Changed

• Prompt cache by @saddam213 in #4
• Supertonic TTS Pipeline by @saddam213 in #7

Full Changelog: v0.1.39...v0.1.81

v0.1.39 - AMD Nitro Pipeline

What's Changed

• AMD Nitro Pipeline
• Llama-3.2-1B pipeline
• ImageTensor GetPixel/SetPixel support
• Video background/foreground masking,removal

New Contributors

• @jdluzen for the support and contribution to #2

Full Changelog: v0.1.18...v0.1.39

v0.1.18 - OpenPose Support

OpenPose Extractor

Features

• Image pose extraction
• Video pose extraction
• Windows demo for image and video

Full Changelog: v0.1.17...v0.1.18

Extractor Demo App

TensorStack Extractor Demo App

The first release of the TensorStack Extractor Demo, a WPF application showcasing the capabilities of TensorStack.Extractors project.

Features

• Simple WPF interface for image and video
• Depth, SoftEdge, Line-Art support
• Background/Foreground Removal
• Background/Foreground Masking

Upscaler Demo App

TensorStack Upscaler Demo App

The first release of the TensorStack Upscaler Demo, a WPF application showcasing the capabilities of TensorStack.Upscaler.

This demo allows users to upscale images and videos using the TensorStack backend, with optional RIFE frame interpolation for smoother playback.

Features

• Simple WPF interface for image and video upscaling
• Comparison view
• Support for RIFE frame interpolation

Image/Video Upscale & Interpolation

TensorStack.Upscaler

The first official release of TensorStack.Upscaler, part of the TensorStack suite.
This library provides high-performance image and video upscaling using ONNX Runtime and supports both buffered inputs (e.g. file-based) and streaming sources (e.g. camera, pipeline, or frame feeds).

Key Features

• Image Upscaling – Enhance image resolution with minimal loss in quality.
• Video Upscaling – Process video frames from file or stream sources with synchronized output.
• RIFE Frame Interpolation – Integrated RIFE model for smooth motion and higher effective frame rates.
• Flexible Input Handling – Supports buffered and real-time processing pipelines.
• TensorStack Integration – Built to work seamlessly alongside other TensorStack components like TensorStack.Image and TensorStack.Video.

Technical Notes

• Built on ONNX Runtime with multi-provider support (CPU, CUDA, DirectML).
• Compatible with TensorStack.Image & TensorStack.Video tensors.

Known Limitations

• Interpolation and upscale models currently run sequentially per frame (parallel batching planned).
• Limited metadata passthrough for streaming sources.

Docs & Examples

TensorStack.Upscaler Readme

Nuget Packages

https://www.nuget.org/packages/TensorStack.Upscaler
https://www.nuget.org/packages/TensorStack.Common
https://www.nuget.org/packages/TensorStack.Providers.CPU
https://www.nuget.org/packages/TensorStack.Providers.DML
https://www.nuget.org/packages/TensorStack.Providers.CUDA
https://www.nuget.org/packages/TensorStack.Image
https://www.nuget.org/packages/TensorStack.Image.Bitmap
https://www.nuget.org/packages/TensorStack.Image.BitmapImage
https://www.nuget.org/packages/TensorStack.Image.ImageSharp
https://www.nuget.org/packages/TensorStack.Video
https://www.nuget.org/packages/TensorStack.Video.Windows