Awesome

Model Tilt for Comfy UI

What is this?

"Model tilt" is a process of adding noise to a subset of the model weights to produce potentially interesting results. I see this as a parallel to cartrigde tilting, hence the name.

How do I use this?

• Copy model_tilt.py to custom nodes directory in Comfy UI
• Apply the patch using git am model_patcher_add_tilt.patch

If done correctly, you will now see a "Model Tilt" node available under "model_patches". This node takes a model as an input and outputs a model with applied noise.

You can tune the following parameters:

• target: a substring of a model key; all the weights which names include this substring will be affected. I personally like to tilt 1.norm or 1.attn.
• tilt_strength: amount of noise to add; the noise will be generated in the [-tilt_strength .. tilt_strength] range.
• tilt_seed: seed to use for tilt noise generation; each subsequent weight tensor will have this seed incremented internally.
• mask_strength: threshold for the mask; the formula is basically weight = weight + noise * mask, where mask is a noise tensor with values less than 1 - mask_strength mapped to 0 and other values mapped to 1.
• mask_seed: seed to use for mask noise generation; each subsequent weight tensor will have this seed incremented internally.

Can you show me some examples?

Workflow example