Resin is a ML framework and search engine of vector areas with hardware accelerated vector operations
Resin comes pre-loaded with two vector set aside configurations (
devices), one obtain-of-words
IModel implementation for textual lisp
and one IModel implementation for MNIST photos.
You would accelerate in your have faith devices. You attain so by enforcing IModel, whose predominant characteristic is to advise to Resin
You customise Resin to your needs by plugging in your have faith practising algorithms into the indexing pipelines.
The artefact of a indexing session is a traversable, scannable and deployable index that you just would possibly grasp interplay with thru
the Resin web GUI, its learn/write
JSON HTTP API, or programmatically thru the
You would populate Resin alongside with your recordsdata, ask it and in diverse ways grasp interplay with it by the utilization of
the built-in web search GUI, otherwise that you just would be in a position to:
- manufacture personalized-made instructions (
ICommand) and accomplish them thru the commandline utility DbUtil.exe
- write recordsdata by HTTP POST-ing JSON formatted recordsdata to the built-in HTTP server write endpoints, and ask by HTTP GET-ing
- write IModel implementations
- programatically scan, traverse, assign calculations over and in diverse ways manipulate your indices.
- Sir.HttpServer: HTTP search provider with HTML GUI and HTTP JSON API for reading and writing.
- Sir.DbUtil: Executes instructions that enforce
Sir.ICommand. Write, validate, ask and more thru expose-line.
ICommandimplementations for downloading and indexing Total Scoot WAT and WET recordsdata.
ICommandimplementations for practising and trying out the accuracy of a index of MNIST photos.
- Sir.Search: In-process search engine with kinds for reading (querying) and writing as effectively as two IModel implementations (
- Sir.Core: Interfaces and kinds that have to clean be shared across libraries and apps, such because the
- v0.1a – obtain-of-characters vector set aside language model
- v0.2a – HTTP API
- v0.3a – ask language
- v0.4 – linear classifier listing model
- v0.5 – semantic language model
- v1.0 – disclose model
- v2.0 – listing-to-disclose
- v2.1 – disclose-to-textual lisp
- v2.2 – textual lisp-to-listing
- v3.0 – AI