Skip to content

2.0.11

Choose a tag to compare

View all tags
@Zalunda Zalunda released this 20 Nov 20:34
· 113 commits to main since this release
2.0.11
51c7649

2.0.9:

  • SubtitleVerb: Merged TranscriberAudioSingleVADAI and TranscriberImageAI into a "multimodal" TranscriberAI that can receive audio and/or image. This allow to add the image to singevad-ai-refined to get a much better TranslationAnalysis that contain targeted image context.
  • SubtitleVerb: Added the possibility to 'inject' context only binary in the request. For example, images taken from the 'gap' between the VOD.
  • SubtitleVerb: Fixed bug where my 'json fixing' code could add a ',' inside the value, like a translated text (ex. "Say \"Please\"" was replaced by "Say \",Please\") because I didn't ignore quote inside a string value.
  • SubtitleVerb: Changed the response handling if the StartTime received from the AI didn't match any of the node's StartTime. Now, I'll match the closest node, if it less then 50 ms away. If not, the same exception as before is raised.
  • SubtitleVerb: Adjustement to use gemini-3-pro-preview instead of 2.5.
  • SubtitleVerb: Default configuration for ManualHQWorkflow is using Gemini 3.0 (preview) for all the steps.

Full Changelog: 2.0.8...2.0.9

2.0.10:

Reverted to Gemini 2.5 pro for full-ai. The timings with 3.0 preview are really bad.

Full Changelog: 2.0.9...2.0.10

2.0.11:

  • SubtitleVerb: Fixed so that an empty JSON array (i.e. "[]") is considered a valid JSON. It was not before.
  • SubtitleVerb: Re-added an explicit rule to break subtitles on characters like 。！？in 'full-ai' step.
  • SubtitleVerb: Allowed more format when parsing TimeSpan (i.e. accept 1.200, 1.20 or 1.2). Before, it only accepted 3 numbers after the dot.

Full Changelog: 2.0.10...2.0.11

Assets 3
Loading