OCI Data Science
a cloud service focused on serving data scientists through the full ML cycle
OCI Data Science Accelerated
methodology of allowing data scientist to work as efficiently as possible
OCI Data Science Collaborative
a system in which sharing and reproducing models is effective
OCI Data Science Enterprise-Grade
OCI provides a full-managed system for data scientists
OCI Data Science Projects
are containers that hold/represent workspace w/ assets such as notebook sessions and models
Tenancy
is the overhead controller for all in OCI
Notebook Sessions
a Juypter lab environment w/ preinstalled libraries & configurable cloud hardware
Conda Environment
Open-source environment and package management for python that helps with (install/run/update) packages and/or dependencies
Accelerated Data Science (ADS)
a python library that automates/simplifies the data science workflow
Model Deployments
allows you to deploy models in the model catalog and expose w/ HTTP endpoints
Jobs
are predefined repeatable ML tasks within the infrastructure
Remote Direct Access Memory (RDMA)
a technology that allows data to be directly transferred between memory of two computers bypassing the CPU and OS
OCI Super clusters
a huge system of GPU’s modified to utilize the RDMA technology
Clos Network
a multistage circuit-switching network that uses a lot of smaller branches connected to GPU s to allow for communication between them all
Network Locality Hints
basically the network endpoints of a GPU are exposed in a way that you can choose the pathways your block connections are for latency
Oracle Select AI
allows you to query databases w/ natural language instead of SQL
Oracle Vector Search
conducts similarity semantic searches by storing/querying high dimensional vector representations
OCI AI Services Categories
Digital Assistant, Language, Vision, Speech, Document Understanding
Data Type Store on Oracle Database 23ai
Vector
OCI Language
pretrained models to perform language analysis; Detect language, Identifiers entities and sentiment, key phrases, and can classify general topic of the language
OCI Speech
transcribes audio/video, processes data directly into object storage, and can generate timestamped accurate transcriptions
OCI Vision
can detect objects inside of an image; Can label a scene
OCI Vision Document AI
can extract text from images, can parse documents into 10 different types, can detect the language, extract tables, and find key value text
OCI Speech Accurate Transcriptions
produces an accurate and easy to use JSON and SubRip Subtitle (SRT)