instill-core
instill-core copied to clipboard
instill-ai
🔮 Instill Core is a full-stack AI infrastructure tool for data, model and pipeline orchestration, designed to streamline every aspect of building versatile AI-first applications
**Why** I suggest we make the error message consistent across all backends, so the frontend can present more constructive error messages to the users. **Now** Right now, our API error...
User may have local environment port conflict with VDP backends' default ports. VDP needs to allow ports configuration in `.env`.
We should use Temporal workflow for the following operations - model creation - model deployment - pipeline creation Related issue: #130
At the moment, a pipeline in **async** mode still relies on **sync** model inference and only asynchronously write to the destination. We should make the `model-backend` trigger endpoints use Temporal...
- [x] gRPC - [x] HTTP - [x] MySQL - [x] Postgres - [x] Local CSV - [x] Local JSON
Add clear contribution guideline or template for - [ ] General contribution - [ ] Contribute a model - [ ] Contribute a demo - [ ] Contribute a connector
Translate README into different languages - [ ] [简体中文](https://github.com/instill-ai/vdp/blob/main/README_zh-hans.md) - [ ] [繁體中文](https://github.com/instill-ai/vdp/blob/main/README_zh-hant.md)
Currently, only pipeline-backend supports search or filter endpoints ([AIP-160](https://github.com/instill-ai/vdp/issues/AIP-160)) (See https://github.com/instill-ai/vdp/issues/109). We need to implement the same function for all backends - [ ] Connector-backend @pinglin - [ ] Model-backend...
Similar to #109, if a resource (source/model-instance/destination) is online - if the resource is used by any **online** pipeline, we should not allow the operation to change the resource state...
Metadata
Owner
Metadata
🔮 Instill Core is a full-stack AI infrastructure tool for data, model and pipeline orchestration, designed to streamline every aspect of building versatile AI-first applications