SwanLab

Model Serving & MLOps

A professional, modern AI training analysis platform that integrates with over 50 major ML frameworks and easily connects to experiment code

Python Latest v0.7.19 · 16d ago Security brief →

Features

Supports >50 leading AI training frameworks for seamless integration
Provides real-time visualization and detailed logging of experiment metrics
Offers hardware monitoring across diverse GPU platforms (e.g., AMD ROCm, Huawei Ascend)
Enables collaborative tracking with project pinning, baseline comparison, and webhook alerts

View all 34 releases →

No immediate action

v0.7.18 New feature 22d

wb converter

Open

v0.7.17 Bug fix 28d

Fixed wandb local converter bug.

Full changelog

Tiny Patch

fix wandb local converter by @Zeyi-Lin in https://github.com/SwanHubX/SwanLab/pull/1596

Full Changelog: https://github.com/SwanHubX/SwanLab/compare/v0.7.16...v0.7.17

v0.7.16 New feature 1mo

Notable features

Full changelog

fix/cancel metric dup warning by @Nexisato in https://github.com/SwanHubX/SwanLab/pull/1544
feat/support section control by @Nexisato in https://github.com/SwanHubX/SwanLab/pull/1559

Full Changelog: https://github.com/SwanHubX/SwanLab/compare/v0.7.15...v0.7.16

v0.7.15 New feature 1mo

Notable features

Full changelog

feat: swanlab.save() file uploader by @Nexisato in https://github.com/SwanHubX/SwanLab/pull/1515
fix: hide overwrite log by @Nexisato in https://github.com/SwanHubX/SwanLab/pull/1535

Full Changelog: https://github.com/SwanHubX/SwanLab/compare/v0.7.14...v0.7.15

v0.7.14 Bug fix 2mo

Fixed file uploads above 100MB silently dropping.

Full changelog