Enhancements - Jan 2025 update

[techthoughts2]

Pull Request

Description

• Module changes:
  • Added two new properties to all model documentation objects:
    • Image - designates if the model can generate images
    • Video - designates if the model can generate video
    • Vision - re-purposed to now designate if the model can see and describe uploaded content
      • This required some logic re-work in several functions that were previously using Vision to designate image capability
  • Added inference profile ID support to all models IAW Supported cross-region inference profiles
    • pwshBedrock will now properly format the inference profile ID for all models that support which enables the ability to make Bedrock calls to other regions
  • Reset-ModelContext - fixed issue where certain models were missing from validation.
  • Amazon
    • Invoke-AmazonNovaTextModel - added support for new Nova Text models: amazon.nova-pro-v1:0, amazon.nova-lite-v1:0, amazon.nova-micro-v1:0
    • Invoke-AmazonImageModel
      • added support for new Nova image model: amazon.nova-canvas-v1:0
      • fixed bug where files would not always save to drive
      • CfgScale parameter maximum adjusted from 10 to 9.9
  • Meta
    • Removed meta.llama2-13b-chat-v1 and meta.llama2-70b-chat-v1 from all functions as Bedrock has EOL these two models. - Breaking Change
    • Updated Invoke-MetaModel to support new model: meta.llama3-3-70b-instruct-v1:0
  • stability.ai
    • Updated Invoke-StabilityAIImageModel to support new model: stability.sd3-5-large-v1:0
  • Updated pricing for all models
• Build changes:
  • Updated bootstrap file to latest versions of modules

License

By submitting this pull request, I confirm that my contribution is made under the terms of the projects associated license.