You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Why isn't the 'control_after_generate' parameter defined in NODE's INPUT_TYPES like other parameters, and then handle the seed in Python code? Also, are there any other parameters in the 'widgets_values' of the 'workflow.json' file that are processed in the frontend?If so, is there a unified place to manage them?
The text was updated successfully, but these errors were encountered:
It seems that "widgets_values" cannot correspond one-to-one with the INPUT_TYPES of the node. Is the issue of unmatchable 'control_after_generate' parameters only present in KSampler, KSamplerAdvanced, and PrimitiveNode?
It seems that "widgets_values" cannot correspond one-to-one with the INPUT_TYPES of the node. Is the issue of unmatchable 'control_after_generate' parameters only present in KSampler, KSamplerAdvanced, and PrimitiveNode?
It's not just a node problem. Actually control_after_generate is a virtual subwidge of widget which is named seed or noise_seed.
And the PrimtivieNode is a special case.
It seems like ComfyUI offers a method to save the workflow in API format. Is there still a need for the functionality of Graph2Prompt to support the use of non-API format workflows through API calls in Python? I've already completed a demo for this feature. If it's still needed, I can submit a pull request for it.
Why isn't the 'control_after_generate' parameter defined in NODE's INPUT_TYPES like other parameters, and then handle the seed in Python code? Also, are there any other parameters in the 'widgets_values' of the 'workflow.json' file that are processed in the frontend?If so, is there a unified place to manage them?
The text was updated successfully, but these errors were encountered: