Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Feature] [VDP] [Components] Connects AI vendors #1024

Open
1 task done
chuang8511 opened this issue Jun 20, 2024 · 2 comments
Open
1 task done

[Feature] [VDP] [Components] Connects AI vendors #1024

chuang8511 opened this issue Jun 20, 2024 · 2 comments
Labels
feature New feature or request need-triage Need to be investigated further

Comments

@chuang8511
Copy link
Member

chuang8511 commented Jun 20, 2024

Is There an Existing Issue for This?

  • I have searched the existing issues

Where do you intend to apply this feature?

Instill Core, Instill Cloud

Is your Proposal Related to a Problem?

Background

We want users to have more optional models to be used

Describe Your Proposed Solution

User Stories

  • As a user, they can easily trigger a pipeline to use a model in the popular vendors.

Highlight the Benefits

The users can use more and more models in VDP

Anything Else?

Planned Components

Milestones

  1. Please create the skeleton PR first for the incoming components
  2. Connect those components.
  3. Build the designed pipeline after you connect those components.
@chuang8511 chuang8511 added need-triage Need to be investigated further feature New feature or request labels Jun 20, 2024
Copy link

linear bot commented Jun 20, 2024

@namwoam
Copy link
Contributor

namwoam commented Jul 9, 2024

We can differentiate the AI vendors into the following three category:
PaaS (Platform as a service) Providers, which requires complex IAM authentication and project permission settings.

  1. GCP (Google)
  2. AWS (Amazon)
  3. Azure (Microsoft)
  4. Watsonx (IBM)
  5. Oracle Cloud Infrastructure
    MaaS (Model as a service) Providers, which often use simple API keys and serverless model endpoint.
  6. OpenAI - Proprietary text foundation models & image generation model
  7. Anthropic - Proprietary text foundation models
  8. Cohere - Proprietary text foundation models specialized on RAG usecase
  9. Mistral AI - Proprietary & OSS text foundation models
  10. Fireworks AI link - Serverless hosting for OSS models
  11. AI21 labs link - Proprietary text foundation models & task-specific models
  12. Upstage link - Proprietary text foundation models specialized in Korean
  13. Voyage AI link - Proprietary domain-specific text embedding model
  14. Deep infra link - Serverless hosting for OSS models
  15. Deepgram link - Proprietary audio <-> text models
  16. unstructured link - Data Pre-processing
  17. Jina AI link - Proprietary text embedding model & Data Pre-processing
  18. BRIA link - Proprietary image generation model
  19. Nvidai NIM link - Serverless hosting for OSS models
  20. Lamini link - Serverless hosting for OSS models
  21. Together AI link - Serverless hosting for OSS models
    SaaS (Software as a service), which often could only be accessed via GUI, thus impossible to integrate into VDP
    example: OpenArt link

Others:

  1. Ollama link OSS model self-hosting solution.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
feature New feature or request need-triage Need to be investigated further
Projects
Status: 🚧 Planned
Development

No branches or pull requests

2 participants