AnyMetrica - commercial analytical software targeted to work with media sources analysis.
Docker
(>=18
) and node.js
(==v10.13.0
) will be required for SDK installation and they are included as mandatory step as SDK installation process.
$ ./install.sh
$ ./install.sh
> install.bat
Following command will download Android Studio and install 27.0 Version of SDK
$ npm run bootstrap:mobile
-
/anymetrica
- root repo where this doc file should be locatedprotocol buffers defining all interfaces contracts -
/anymetrica-api
- protocol buffers defining all interfaces contracts -
/anymetrica-registry
- core business logic entities and their relations registry service -
/anymetrica-scope
- administrative Web UI -
/anymetrica-sdk
- SDK installer -
/anymetrica-sosalka
- ML models and media assets downloader/installer/manager -
/deployment
- deployment scripts -
/dockerfiles
- Docker files -
/scripts
- misc scripts
/install.sh
- Linux/MacOS/POSIX SDK installation/install.bat
- Windows NT 64 installation
Execute $ yarn run
command in repo root to see all launchers that could be used as: $ yarn run [SCRIPT NAME]
.
/DATA
- Local ML data folder not included in repo/build
- Local build data folder not included in repo/logs
- Local logs not included in repo
-
grid-grpc - provides GRPC bridge and code generations for the generic web/ws clients.
-
PyTorch Speaker Verification (forked from) - PyTorch implementation of speech embedding net and loss described here: Generalized end-to-end loss for speaker verification (2018).
Also contains code to create embeddings compatible as input for the speaker diarization model found at Google UIS RNN.
-
protobuf-jsonschema-validate Make JSON schema based on protoc-gen-validate format.