digital-garbageandClaude Opus 5
ab9cbdb556
Ported from v4, minus two mechanisms that are no longer needed. The LoRA loader takes one LoRA and chains, so the FlexibleOptionalInputType dict subclass that lied to ComfyUI's validator has nothing left to do. The CLIP loader uses numbered slots like core's DualCLIPLoader instead of a JSON array in a string widget. Autogrow looked like the right answer for the encoder list and is not: it sets force_input on any widget input, so a list of combos renders as sockets and the file picker stops being a picker. Every core use of it is a connection type. GGUF and SLA are deliberately not here. Both are capabilities rather than loader options, and both belong in modules of their own. The tests follow what the loaders PRODUCE rather than stopping at the call: the attention override is carried into comfy's real dispatch and checked against every backend this machine registers, because an override that only fails at the first attention call fails at sampling, not at load. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
59 lines
2.7 KiB
Python
59 lines
2.7 KiB
Python
"""The LoRA loader.
|
|
|
|
One LoRA per node, chained for more. v4 carried N LoRAs in a single node behind a
|
|
JSON list widget and a `FlexibleOptionalInputType` -- a dict subclass that lied to
|
|
ComfyUI's validator so arbitrary keys would pass. Chaining needs neither: it is
|
|
what ComfyUI's own LoraLoader does, every strength is a real socket, and the graph
|
|
shows how many LoRAs are actually applied instead of hiding the count in a string.
|
|
"""
|
|
|
|
import comfy.sd
|
|
import comfy.utils
|
|
import folder_paths
|
|
from comfy_api.latest import io
|
|
|
|
|
|
class FunPackLoraLoader(io.ComfyNode):
|
|
@classmethod
|
|
def define_schema(cls) -> io.Schema:
|
|
return io.Schema(
|
|
node_id="FunPackLoraLoader",
|
|
display_name="FunPack LoRA Loader",
|
|
category="FunPack/Loaders",
|
|
description="Apply one LoRA to a model, and optionally its text encoder. "
|
|
"Chain these for more than one.",
|
|
inputs=[
|
|
io.Model.Input("model"),
|
|
io.Combo.Input("lora_name", options=folder_paths.get_filename_list("loras")),
|
|
io.Float.Input("strength_model", default=1.0, min=-100.0, max=100.0, step=0.01),
|
|
io.Clip.Input("clip", optional=True,
|
|
tooltip="Leave unwired for a model-only LoRA."),
|
|
io.Float.Input("strength_clip", default=1.0, min=-100.0, max=100.0, step=0.01,
|
|
optional=True),
|
|
],
|
|
outputs=[
|
|
io.Model.Output(display_name="model"),
|
|
io.Clip.Output(display_name="clip"),
|
|
io.String.Output(display_name="status"),
|
|
],
|
|
)
|
|
|
|
@classmethod
|
|
def execute(cls, model, lora_name: str, strength_model: float,
|
|
clip=None, strength_clip: float = 1.0) -> io.NodeOutput:
|
|
# Doing nothing is a valid request, but it must return the SAME objects
|
|
# rather than clones: a clone here would silently drop patches a later
|
|
# loader in the chain applied to the original.
|
|
if strength_model == 0 and (clip is None or strength_clip == 0):
|
|
return io.NodeOutput(model, clip, f"{lora_name}: both strengths are 0, not applied")
|
|
|
|
path = folder_paths.get_full_path_or_raise("loras", lora_name)
|
|
lora = comfy.utils.load_torch_file(path, safe_load=True)
|
|
patched_model, patched_clip = comfy.sd.load_lora_for_models(
|
|
model, clip, lora, strength_model, strength_clip)
|
|
|
|
applied = [f"model={strength_model}"]
|
|
applied.append(f"clip={strength_clip}" if clip is not None else "clip=not wired")
|
|
return io.NodeOutput(patched_model, patched_clip,
|
|
f"FunPack LoRA Loader | {lora_name} ({', '.join(applied)})")
|