tanishqgandhi1908 opened a new issue, #6494: URL: https://github.com/apache/texera/issues/6494
## Summary Enable users to bring their own trained ML models into Texera and use them inside workflows - upload a model, keep versions of it, share it, and run inference on workflow data through it. Today users can bring data (datasets) but not models; this closes that gap so an end-to-end ML workflow can live entirely in Texera. ## User stories - As a user, I can **upload a trained model** and its files to Texera. - As a user, I can keep **multiple versions** of a model and pick which one to use. - As a user, I can **browse, search, and manage** my models, and share them with others (private / public, like other resources). - As a workflow author, I can **select a model** and **run inference** on my data with it. ## Scope - **MVP:** PyTorch models. Upload → version → manage → use in a UDF workflow for inference. - **Later:** more frameworks, a standardized model format, and a no-code inference operator ## Approach (high level) Models are a new first-class resource in Texera with their own storage, versioning, and access control, and a way to reference a model from a workflow operator. Detailed design lives in the sub-tasks below. ## Sub-tasks - [ ] Resource-type namespacing for asset paths - [ ] Model metadata storage - [ ] Model file storage & path resolution - [ ] Model management API (upload / version / access) - [ ] Model management UI - [ ] Use a model inside a workflow - [ ] Sharing, discovery & docs -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
