Lava Deep Learning 0.2.0

[mgkwill]

The lava-dl library version 0.2.0 now supports automated generation of Lava processes for a trained network described by hdf5 network configuration using our Network Exchange (NetX) library.

New Features and Improvements

• Released Network Exchange (NetX) library to support automated creation of Lava process for a deep network. We support hdf5 network exchange format. Support for more formats will be introduced in future. (PR #30, Issue #29)

Bug Fixes and Other Changes

• Fixed bug with pre-hook quantization function on conv blocks (PR#13)

Breaking Changes

• No breaking changes in this release

Known Issues

• Issue training with GPU for lava-dl-slayer on Windows machine.

What's Changed

• Create PULL_REQUEST_TEMPLATE.md & ISSUE_TEMPLATE.md by @mgkwill in Create PULL_REQUEST_TEMPLATE.md & ISSUE_TEMPLATE.md #27
• Hardware neuron parameters exchange and fixed precision instruction precision compatibility by @bamsumit in Hardware neuron parameters exchange and fixed precision instruction precision compatibility #25
• Pilotnet link fix by @bamsumit in Pilotnet link fix #31
• Bugfix: CUBA neuron normalization applied to current state by @bamsumit in Bugfix: CUBA neuron normalization applied to current state #35
• Netx by @bamsumit in Netx #30
• Streamline PilotNet SNN notebook with RefPorts by @bamsumit in Streamline PilotNet SNN notebook with RefPorts #37
• Fix for failing tests/lava/lib/dl/netx/test_hdf5.py by @bamsumit in Fix for failing tests/lava/lib/dl/netx/test_hdf5.py #44
• Update ci-build.yml by @mgkwill in Update ci-build.yml #42
• Install by @mgkwill in Install #45
• Lava Deep Learning 0.2.0 by @mgkwill in Lava Deep Learning 0.2.0 #46
• Lava Deep Learning 0.2.0 - update lock by @mgkwill in Lava Deep Learning 0.2.0 - update lock #47

Full Changelog: v0.1.1...v0.2.0

---

This discussion was created from the release Lava Deep Learning 0.2.0.