1.7.0

Update v1.7.0

Bug fixes

• Fix UI breaking on latest gradio (fixed by @hounddogie)
• Fix dynamic VRAM reallocation (reported and fixed by @phr00t, issue #17)
• Attempt to fix incompatibility with other extensions which modify output too (e.g. Alltalk TTS, reported by @guispfilho, issue #18)

Misc

• JSON logits: switch to an implementation based on Outlines (used by tool mode via tool_mode_force_json_output_enabled). Still WIP and probably not working correctly. Experimental and not recommended yet.

Special thanks to @phr00t and @hounddogie for their contributions.

DOWNLOAD AND INSTALLATION INSTRUCTIONS HERE

1.6.1

Features

• Added new tool / function call based generation mode.

This new mode uses tool / function calls to give the LLM model the ability to trigger and generate images itself!
Very similar to how GPT4 calls Dall-E when you request image generation.

Works with no hardcoded triggers, has multilingual support, and uses past chat history context. For example, if you told it earlier to only generate cartoon style images, it will remember this and do it until you say otherwise.

This mode does not support chat streaming and requires special models (like Command-R or Llama-2-chat-hf-function-calls-v2) and some minor system prompts adjustments to describe the tools / function calls available. If you do not use a transformers / HF based loader, you will probably also have to modify start_with so it always outputs JSON. Consult the sample settings.yaml and the model documentation for more information on this.

Bug fixes

• Added missing "None" option for VAEs
• Fixed disable streaming during image generation not working.
• Fixed debug mode not working correctly
• Replace FaceID checkboxes with dropdown as SD.Next does not support multiple FaceID anymore
• Removed invalid gradio calls such as refresh_button.update(...)
• Fixed network requests spam because of wrong gradio usage, this should fix performance issues on the UI side

Misc

• Improved prompt normalization by removing newlines, hashtag, etc.
• Replaced mypy with pyright / pylance

DOWNLOAD AND INSTALLATION INSTRUCTIONS HERE

Release v1.4

• Add support for FaceID feature of SD.Next (currently only available on SD.Next "dev" branch)
• Add support for IP Adapters feature of SD.Next (currently only available on SD.Next "dev" branch)
• Add Hires.fix related params
• Update default generation params to what RealisiticVision v6.0_B1 recommends
• Rename "enhace_face" related params to "restore_face"
• Fix some params not updating via UI, especially params set to default value and params of type boolean or enum
• Fix unhandled error when downloading or loading images for face swap fails
• Fix UI wrongly setting username as endpoint
• Fix UI wrongly setting password as endpoint
• Fix some AUTOMATIC1111 related API incompatibilities
• Fix trigger on output message, fixes #11
• Fix multithreading issues related to context, fixes #13
• Fix full_quality param not set for SD.Next
• Fix missing trace log when errors are logged

Thanks to @TheInvisibleMage for reporting some of the issues.

Some configuration option defaults have changed. I suggest reconfiguring the settings.yaml from scratch.

DOWNLOAD AND INSTALLATION INSTRUCTIONS HERE

1.3

• Update to latest SD.Next API which is now more aligned with AUTOMATIC1111 (also improves AUTOMATIC1111 compatibility).
  You must be running latest SD.Next or you will get errors about refresh-vaes!
• Added new input_sentence/output_sentence rule matches to only match on a sentence instead of the whole text.
• Added new rule actions for dynamically configuring face swapping.
• Added tag deduplication when multiple rules add the same tags to avoid overemphasizing on them.
• Fixed different implementations for samplers and upscalers API (improves AUTOMATIC1111 compatibility).
• Temporary fix for images no longer generating after clicking on "Continue".
• Fixed various other bugs.

Some configuration option defaults have changed. I suggest reconfiguring the settings.yaml from scratch.

DOWNLOAD AND INSTALLATION INSTRUCTIONS HERE

Release v1.2

• Added image generation rules:
  You can now adjust the image generation prompt with regex-based rules.
  In other words, you can dynamically include tags or lora etc. based on context such as the current character, specific words or phrases occuring (or not occurring) in input or output messages, etc.
• Added an option to trigger image generation based on output message too for interactive mode.
• Updated default input trigger regex pattern. This improves triggering image generation in interactive mode.
• Updated default subject regex pattern. This improves detection of the subject for image generation in interactive mode.
• Fixed various issues with handling HTML messages.
• Removed default_prompt option.

Some configuration options have been renamed. I suggest reconfiguring the settings.yaml from scratch.

DOWNLOAD AND INSTALLATION INSTRUCTIONS HERE

Release v1.1

• Added prompt generation modes
• Added face swap support using ReActor extension
• Added new configuration options for FaceSwapLab extension
• Added debug mode and changed various messages to only log when it's enabled
• Updated readme to add better installation instructions
• Renamed some configuration options to better reflect what they do
• Fixed FaceSwapLab settings never being actually used (except for source face)
• Fixed various options being defined as integer instead of float such as cfg scale or upscaler scale
• Fixed special html characters rendering incorrectly such as apostrophes
• Disabled image generation if not used in chat mode
• Lots of other bugs fixed

DOWNLOAD AND INSTALLATION INSTRUCTIONS HERE