v0.2.0

What's Changed

• fix(ops): Use correct C types for torch/tf Buffer by @papazof in #18
• ci(generate-api-reference): Fix printing step for larger logs by @papazof in #19
• feat: Add vaccel_config and boostrap/cleanup wrappers by @papazof in #20
• refactor: Improve null ptr handling and object representations by @papazof in #21
• feat(resource): Enable the use of multiple paths as input by @papazof in #22
• refactor(ops/exec): Enable calling functions with no input/output by @papazof in #23
• feat: Add helpers for NumPy and PyTorch conversions by @papazof in #24
• refactor(version): Bump vAccel version by @ananos in #25

Full Changelog: v0.1.0...v0.2.0

v0.1.0

v0.1.0 · Initial Release

This is the first release of the vAccel Python bindings, enabling Python developers to use vAccel for accelerated operations like image inference, tensor processing, and generic op execution.

Features

• Noop Operation Example: Basic example demonstrating the vAccel noop (#2)
• GenOp Support: Introduced bindings for the genop interface (#3)
• Image Inference API: Python bindings for image classification/inference using genop (#4, #5)
• Vector Operations: Added support for vector-related operations with docstrings (#13)
• Exec and Resource Binding: Exposed exec, exec_with_resource, and added Object class (#14)

Maintenance & Enhancements

• Packaged Python bindings for distribution (#7)
• Added versioning and improved setup handling (#10)
• Polished TensorFlow (TF) bindings (#8)
• Updated README and added a user guide (#11)
• Repository cleanup and major refactors for alignment with latest vAccel updates (#1, #15, #16)

Full Changelog: https://github.com/nubificus/vaccel-python/commits/v0.1.0