Add LoRA loader nodes for quantized models#31
Open
FearL0rd wants to merge 1 commit into
Open
Conversation
Add QuantizedLoRALoader and QuantizedLoraStack nodes that use bypass mode to properly handle QuantizedTensor wrappers and all quantization formats. Supports stacking multiple LoRAs with individual strength controls and on/off toggles. Key features: - QuantizedLoRALoader: Load single LoRA with bypass mode - QuantizedLoraStack: Stack up to 8 LoRAs with on/off toggles - Uses comfy.sd.load_bypass_lora_for_models() for quantized compatibility - Supports all quantization formats (INT8, FP8, MXFP8, NVFP4, BNB 4-bit) - Bypass mode injects LoRA computation during forward pass
Owner
|
There is already bypass mode LoRA loaders in ComfyUI core though. |
Author
Try to stack 2 loras using the bypass |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Add QuantizedLoRALoader and QuantizedLoraStack nodes that use bypass mode to properly handle QuantizedTensor wrappers and all quantization formats. Supports stacking multiple LoRAs with individual strength controls and on/off toggles.
Key features: