v1.3.0 Roadmap

[FeelyChau]

Here is the list we need to do in v1.3.0:

1. Pipeline & plugin @WenheLI

• dataset management. Data Collect Plugin re-use #569
• configurable params for jobs in one pipeline. Configurable parameters per job #565
• other plugins need to support for stream-based pipeline. Stream-based Pipeline & Plugins #571

2. Daemon @FeelyChau

• plugin package manager.
• support specific version for plugins pipeline: support specific version for plugins #448.
• test cases for services. daemon: test cases for services #564
• package source management.Daemon should handle npm packages from different sources #579

3. Pipboard @yorkie

• integration with tensorboard(tensorboard integration imgcook/pipboard#4)
• dataset visualization(feature: dataset visualization imgcook/pipboard#6)
• model visualization(feature: model visualization imgcook/pipboard#5)

4. CI/CD @FeelyChau

• auto-benchmark & Visualization on building/executation/installing speed per version Benchmarking & Visualization #572
• pipelines test on self-hosted(GPU). CI/CD: pipelines test on self-hosted workflow. #567
• Fill and automate smoke test cases. tests: Fill and automate smoke test cases. #570

5. Models @rickycao-qy

• models for structured data (svm, gbdt ...) plugins for structured data and relative models #566
• YOLO / SSD (current od is too slow and no model for simple data) To provide faster object-detection models, such as SSD and YOLO #568
• WASM executable model meta: Wasm Convertor Package #563

6. Bugs

[yorkie]

What's the progress of our smoke testing in v1.3.0?

[yorkie]

The label versioning is used for release, added a new label roadmap :)

[yorkie]

@FeelyChau any updates about our 1.3 release?

[FeelyChau]

@FeelyChau any updates about our 1.3 release?

How about 2022-12-15?

[FeelyChau]

v1.3.0 has been released, see https://github.com/alibaba/pipcook/tree/v1.3.0

[FeelyChau]

Features not implemented on v1.3.0 will continue to iterate on the 2.x branch.