daytona Releases
99 releases of daytonaio/daytona
- v2.33.0
Updates dependencies to Electron v38, refactors test suites using Jest, and fixes bugs in calendar and weather modules. Requires Node.js v22.18.0+.
Sep 30, 2025
- v0.3.0v0.3.0 is now live ! 🚀
Adds eraser functionality, multi-language support (i18n), tutorial animations, and security fixes to the drawing board application.
Sep 14, 2025
- v3.2.0PaddlePaddle 3.2.0 Release Note
Enhances large model training with FlashMask V3 for attention efficiency, dynamic adaptive memory offloading, and a new CUDA-like operator registration mechanism. Improves API compatibility with Safetensors and industry-standard practices.
Sep 8, 2025
- v15.2
Adds support for loading proxy configuration rules via files, base64 strings, or direct strings to simplify proxy parameter management.
Sep 2, 2025
- v3.1.1PaddlePaddle 3.1.1 Release Note
Focuses on stability for large models by fixing numerical precision and functional issues. Optimizes FP8 quantization and Slice performance. Adds device switching and type conversion capabilities.
Aug 21, 2025
- v2.0.0
Major version update separating engine and client responsibilities, introducing a new Breeze model type, and adding instrumented tests.
Aug 5, 2025
- v3.1.0PaddlePaddle 3.1.0 Release Note
Optimizes the Automatic Parallel architecture for dynamic graph performance, introduces FP8 low-precision training for 10-20% speedup, and lowers hardware adaptation costs with operator reuse mechanisms.
Jun 29, 2025
- v15.1
Jun 11, 2025
- v1.0.6
Bug fix release addressing an issue with LLM configuration setting values exceeding preference default boundaries.
May 16, 2025
- v1.0.5
Bug fix release addressing screen rotation issues and missing imports in the application settings.
May 15, 2025
- v1.0.4
Adds a new LLM model configuration page, enhances UI for TTS triggering with failure handling, and upgrades model switching capabilities.
May 14, 2025
- v0.2.0
This release introduces new trainers for channel-concatenated control conditioning (Wan2.1, CogView4), adds support for FLUX.1-dev and Wan2.1 I2V models, and enhances training infrastructure with multiple attention providers.
Apr 25, 2025
- v0.9.1
Apr 24, 2025
- v15.0
Apr 1, 2025
- v0.1.0
The initial stable release of the library, adding official support for CogView4 and Wan 2.1 models with released checkpoints, and improving data loading capabilities.
Mar 21, 2025
- v0.0.1
The initial work-in-progress release of the library designed to support accessible training of diffusion models, specifically CogVideoX, LTX Video, and Hunyuan Video.
Feb 24, 2025
- v0.1.1-alpha
This release introduces significant new vocoder and dataset support, including DiffWave and Hi-Fi TTS preprocessing, alongside various bug fixes and environment compatibility updates.
Feb 23, 2024
- v0.1.0-alpha
This marks the initial Alpha release of the Amphion repository, establishing the foundational codebase and structure.
Feb 23, 2024
- v3.1fixv3.1
Minimal release notes: "1.优化了内网穿透功能,bridge,client和server只需要启动一个即可。 server端启动的时候可以指定client端要暴露的一个或者多个端口。 2.修复了重复解析命令行参数的问题。 3.手册增加了微信接"
Oct 10, 2017