A fast, easy-to-use, production-ready inference server for computer vision supporting deployment of many popular model architectures and fine-tuned models.
added GH action to ensure our packages (inference aka inference-cpu, inference-gpu, inference-cli and inference-sdk) are installable against grid of supported py version and asserted that one can import the package in Py interpreter / use CLI
fixed workflows notebooks examples regarding changes in new workflows EE
fixed broken links in inference server landing page
thanks to @grzegorz-roboflow for fixing instance-segmentation bug causing fatal errors on batch predictions when one of the output was empty
Type of change
Please delete options that are not relevant.
[ ] Bug fix (non-breaking change which fixes an issue)
[x] New feature (non-breaking change which adds functionality)
[ ] This change requires a documentation update
How has this change been tested, please provide a testcase or example of how you tested the change?
Jetson with Jetpack 5.1 - manual run by @PawelPeczek-Roboflow - 🟢 - confirmed that workflows work under HTTP server, also thanks to @yeldarby we have Jetson runners available - [run]
Jetson with Jetpack 4.6 - 🟢 [run] - turned out that there was release of h5py which requires HDF5-related libs in version >=1.10.4, which is not available for OS running behind nvidia base image - pinned here into latest working version
Jetson with Jetpack 4.5 - hardware not available
Any specific deployment considerations
For example, documentation changes, usability, usage/costs, secrets, etc.
Description
PR with release of version
v0.10.0
. With this PR:inference
akainference-cpu
,inference-gpu
,inference-cli
andinference-sdk
) are installable against grid of supported py version and asserted that one can import the package in Py interpreter / use CLIType of change
Please delete options that are not relevant.
How has this change been tested, please provide a testcase or example of how you tested the change?
h5py
which requires HDF5-related libs in version>=1.10.4
, which is not available for OS running behind nvidia base image - pinned here into latest working versionAny specific deployment considerations
For example, documentation changes, usability, usage/costs, secrets, etc.
Docs