0% found this document useful (0 votes)
10 views37 pages

OCI Generative AI Overview and Demos

The document provides an overview of OCI Generative AI, detailing its services, pretrained models, and fine-tuning capabilities. It includes various demos showcasing chat models, embedding models, and configuration setups, along with discussions on prompt engineering and customizing LLMs. Additionally, it addresses security measures and dedicated AI clusters for optimized performance.

Uploaded by

nagendra2118
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
10 views37 pages

OCI Generative AI Overview and Demos

The document provides an overview of OCI Generative AI, detailing its services, pretrained models, and fine-tuning capabilities. It includes various demos showcasing chat models, embedding models, and configuration setups, along with discussions on prompt engineering and customizing LLMs. Additionally, it addresses security measures and dedicated AI clusters for optimized performance.

Uploaded by

nagendra2118
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

1.

Introduction

2. OCI Generative AI
a. Intro

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
b. OCI Generative AI services Work

c. Pretrained Foundational Models

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
d. Fine Tuning

e. Dedicated AI clusters

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
3. Demo OCI Generative AI
a. Screen1

b. Screen2

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
c. Screen3

d. Screen4

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
4. Chat Models
a. Tokens

b. Pre-Trained Chat Models - Command -r

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
c. Chat Model parameters

d. Preamble Override

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
e. Temperature

f. Parameter – Top k

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
g. Parameter – Top p

h. Frequency & Presence Penalties

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
5. Demo - Chat Models

a. Example 1

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
b. Example 2

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
6. Demo – Generative AI Inference API
a. Screenshot 1

7. Demo – config setup for Generative AI Inference API


a. Screen 1

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
8. Embedding Models
a. Intro

b. Word Embeddings

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
c. Semantic Similarities

d. Sentence Embedding

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
e. Embedding Use Cases

f. Embedding Models In Generative AI

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
g. Embedding Models In Generative AI – Categories

9. Demo – Embedding Models

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
10. Prompt Engineering
a. Prompt and prompt Engineering

Co b. LLM’s as next word predictors


nfi
de
nti
al-
Ora
cle
Int
ern
al
c. Aligning LLM’s to follow instructions

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
d. In Context learning & few shot prompting

e. Prompt Formats

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
f. Advanced Prompting Strategies

11. Customize LLMs with your data


a. Intro

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
b. Technic 1 – In-Context/ few shot prompting

c. Fine Tuning Pre trained model

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
d. Fine Tuning Benefits

e. Retrieval Augmented Generation (RAG)

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
f. Customize LLM’s with your data

g. All 3 use

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
12. Fine Tuning & Interference in OCI Gen AI
a. Intro

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
b. Fine Tuning - Work flow

c. Inference - Work flow

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
d. Dedicated AI Clusters

e. T few Fine tuning

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
f. T few Fine tuning- Process

g. Reduce Inference Cost

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
13. Dedicated AI clusters sizing & pricing
a. Unit Types

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
b. Unit Sizing

c. Pricing Example

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
14. Demo – Dedicated Clusters

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
15. Fine Tuning Configuration
a. Intro

b. T-Few – Parameter

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
c. Fine tune Results

d. Accuracy

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
e. Loss

16. Demo – Fine Tuning & Custom Models

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
17. Demo – Inference using End Point

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
18. OCI Generative AI Security
a. Dedicated GPU & RDMA Network

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
b. Model Endpoints

c. Customer Data & Model Isolation

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al
d. Leverages OCI Security Services

Co
nfi
de
nti
al-
Ora
cle
Int
ern
al

You might also like