v3.1.4

• Update yt-dlp

v3.1.3

What's Changed

• Fixed issue with VTT format (#81)
• Update README.md by @miketwo in #75

New Contributors

• @miketwo made their first contribution in #75

Full Changelog: v3.1.2...v3.1.3

v3.1.2

What's Changed

• Add CPU_THREADS environment to configure the CPU threads used. It defaults to 4. (#28)

Full Changelog: v3.1.1...v3.1.2

v3.1.1

What's Changed

• Use actual filename of uploaded media instead of "subtitles.srt" when downloading transcription by @JonasCz in #69
• Display time in hh:mm:ss format instead of seconds by @JonasCz in #70
• reversed transcription list display by @JonasCz in #72

New Contributors

• @JonasCz made their first contribution in #69

Full Changelog: v3.0.1...v3.1.1

v3.1.0 - large-v3 support

Added support for Whisper Large v3 model

• This release bumps the faster-whisper version to 0.10.0, and adds support for the new OpenAI large-v3 model (#48) .

More info about the model:
openai/whisper#1762

Full Changelog: v3.0.0...v3.1.0

v3 - 🐳 Container Consolidation

Warning

BREAKING CHANGES
You must update your docker-compose.yml file to the latest version available in this repository to ensure compatibility with the newest Whishper release (v3). For detailed guidance, refer to the updated documentation or opt for a fresh installation using the get-whishper.sh script.

GPU Version Users: Please follow these specific upgrade instructions to transition from v2.x to v3.x.

Key Updates in This Release

1. 🐳 Container Consolidation: I've streamlined the architecture by merging the three previous Whishper containers into a single container. Additionally, the whishper-proxy container is now redundant, as its functionality is integrated within the main container. As a result, the nginx.conf file is no longer required, simplifying the docker-compose.yml setup for a more user-friendly experience.

2. 🔄 File Upload Progress Bar: Addressing a common user concern (#20), I've added a progress bar for file uploads. This enhancement eliminates the uncertainty experienced during large file uploads where previously no feedback was provided, potentially giving the impression that Whishper had become unresponsive.

Upcoming Feature Teaser: In the next release, look forward to a highly requested feature – the ability to use local folders as media inputs (#15).

---

Changelog

• Simplified Docker Experience: Whishper has evolved into a unified docker container, requiring only a database container and the optional libretranslate for translations.
• Enhanced User Feedback: The new upload progress feature (#20) now provides real-time feedback during file uploads, a feature much awaited by our users.
• Update translation models: When you add new languages to the LT_LOAD_ONLY variable in the .env file, now LibreTranslate will update and download any new languages. This is achieved by adding LT_UPDATE_MODELS: True by default to the Libretranslate environment in the docker-compose.yml file.
• Additional Tweaks and Enhancements: This release also includes various minor fixes and improvements for an overall smoother experience.

Explore the Full Changelog: Whishper v3.0.0 Changes

v2.1.1-beta

fix: no space left while building

v2.1.0-beta

Merge branch 'main' of github.com:pluja/whishper

v2.0.0 - 🔥 GPU Support

⚠️ This release introduces Breaking Changes ⚠️

• Make sure to get the latest docker-compose.yml and .env file versions from the repository!

---

What's Changed

• [#34] Introduced support for GPU with CUDA: read the docs
• Improved startup speed
• Some internal fixes for frontend, backend and API
• Migrated from yarn to pnpm
• Improved install script for Linux/MacOS
• Bump github.com/gofiber/fiber/v2 from 2.49.2 to 2.50.0 in /backend by @dependabot in #32
• Bump postcss from 8.4.28 to 8.4.31 in /frontend by @dependabot in #33

See Full Changelog here: v1.2.0...v2.0.0

v1.3.7-beta

Full Changelog: v1.3.6-beta...v1.3.7-beta