nodes/olivv-cs-ComfyUI-FunPack
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

FunPackStudio: integrate full LoRA chain internally, remove lora_stack output

Browse Source 
LoRA pipeline is now fully internal to Studio:
1. FunPackApplyLoraWeights reads session weight suggestions for the
   current prompt and builds a stack with adjusted model weights
2. FunPackLoraLoader applies the stack to model and clip
3. refine_v2 receives the LoRA-patched model and applies the attn2
   direction patch on top

Removes the lora_stack return value - MODEL output already contains
everything (LoRA patches + direction injection). External lora_stack
input still accepted to bypass internal management entirely.

Adds mode (ltx2/wan) and per_block settings to the LoRA tab in popup,
stored in loras_config dict in studio_settings.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
digital-garbage
2026-05-16 19:18:41 +03:00
co-authored by Claude Sonnet 4.6
parent ff4ebc1b7a
commit b6edaf6db6
2 changed files with 77 additions and 34 deletions
Download Patch File Download Diff File Expand all files Collapse all files
conditioning.py
+66 -32
View File
@@ -11859,8 +11859,8 @@ class FunPackStudio:
Advisor LLM, LoRA management, and Conditioning Adjust under one UI."""
CATEGORY = "FunPack"
RETURN_TYPES = ("CONDITIONING", "STRING", "STRING", "IMAGE", "STRING", "MODEL", "INT", "FUNPACK_LORA_STACK")
RETURN_NAMES = ("modified_positive", "status", "training_info", "loss_graph", "encoded_prompts", "model", "seed", "lora_stack")
RETURN_TYPES = ("CONDITIONING", "STRING", "STRING", "IMAGE", "STRING", "MODEL", "INT")
RETURN_NAMES = ("modified_positive", "status", "training_info", "loss_graph", "encoded_prompts", "model", "seed")
FUNCTION = "run"
DESCRIPTION = (
"FunPack Studio - all FunPack refinement tools in one node. "
@@ -11868,6 +11868,7 @@ class FunPackStudio:
)
_lora_loader = None
_apply_weights_node = None
@classmethod
def _get_lora_loader(cls):
@@ -11882,6 +11883,19 @@ class FunPackStudio:
cls._lora_loader = None
return cls._lora_loader
@classmethod
def _get_apply_weights_node(cls):
if cls._apply_weights_node is None:
try:
try:
from .model_management import FunPackApplyLoraWeights
except ImportError:
from model_management import FunPackApplyLoraWeights
cls._apply_weights_node = FunPackApplyLoraWeights()
except Exception:
cls._apply_weights_node = None
return cls._apply_weights_node
@classmethod
def INPUT_TYPES(cls):
return {
@@ -11909,7 +11923,7 @@ class FunPackStudio:
"model": ("MODEL", {"tooltip": "Diffusion model. Required for LoRA loading and direction injection."}),
"source_image": ("IMAGE",),
"clip_vision_output": ("CLIP_VISION_OUTPUT",),
"lora_stack": ("FUNPACK_LORA_STACK", {"tooltip": "Optional external LoRA stack. Takes precedence over LoRAs configured inside Studio."}),
"lora_stack": ("FUNPACK_LORA_STACK", {"tooltip": "Optional external LoRA stack. Bypasses Studio's internal LoRA management entirely when connected."}),
"refinement_key_input": ("STRING", {"forceInput": True}),
"user_intent_prompt": ("STRING", {"multiline": True, "default": "", "forceInput": True}),
"feedback_prompt": ("STRING", {"multiline": True, "default": "", "forceInput": True,
@@ -11984,40 +11998,60 @@ class FunPackStudio:
except Exception as e:
print(f"[FunPackStudio] Advisor LLM load failed: {e}")
# --- LoRA stack ---
active_lora_stack = lora_stack
# --- LoRA: ApplyWeights (session suggestions) then Load ---
# 1. Build lora_list from studio settings or use provided external stack directly.
# 2. FunPackApplyLoraWeights reads session weight suggestions for current prompt.
# 3. FunPackLoraLoader applies the stack to model/clip.
# 4. refine_v2 then applies attn2 direction patch on top of the LoRA-patched model.
active_lora_stack = lora_stack # external stack bypasses steps 1-2
studio_loras = settings.get("loras", [])
lora_cfg = settings.get("loras_config", {}) if isinstance(settings.get("loras_config"), dict) else {}
lora_mode = str(lora_cfg.get("mode", "ltx2") or "ltx2")
per_block = bool(lora_cfg.get("per_block", False))
if not active_lora_stack and studio_loras and model is not None:
lora_entries = []
for idx, entry in enumerate(studio_loras if isinstance(studio_loras, list) else []):
if not isinstance(entry, dict):
continue
name = str(entry.get("name", "") or "").strip()
if not name or name == "None":
continue
mw = float(entry.get("model_weight", 1.0) or 1.0)
cw = float(entry.get("clip_weight", 1.0) or 1.0)
lora_entries.append({
"slot": idx,
"name": name,
"type": str(entry.get("type", "general") or "general"),
"base_model_weight": mw,
"base_clip_weight": cw,
"model_weight": mw,
"clip_weight": cw,
})
if lora_entries:
raw_stack = {"refinement_key": key, "loras": lora_entries, "per_block": False}
loader = self._get_lora_loader()
if loader is not None:
valid_loras = [
e for e in (studio_loras if isinstance(studio_loras, list) else [])
if isinstance(e, dict) and str(e.get("name", "") or "").strip() not in ("", "None")
]
if valid_loras:
# Build lora_list in the format FunPackApplyLoraWeights expects
clip_weights = {str(e.get("name", "")): float(e.get("clip_weight", 1.0) or 1.0)
for e in valid_loras}
lora_list_json = json.dumps([
{
"name": str(e.get("name", "")),
"type": str(e.get("type", "general") or "general"),
"strength": float(e.get("model_weight", 1.0) or 1.0),
"on": True,
}
for e in valid_loras
])
apply_node = self._get_apply_weights_node()
if apply_node is not None:
try:
model, clip, active_lora_stack, lora_status = loader.load_loras(
model, raw_stack, clip=clip
effective_key = key or "studio"
active_lora_stack, lora_apply_status = apply_node.apply_lora_weights(
active_prompt, effective_key, lora_mode,
per_block=per_block,
lora_list=lora_list_json,
)
print(f"[FunPackStudio] {lora_status}")
# Inject clip_weight values (ApplyLoraWeights only handles model weights)
for entry in active_lora_stack.get("loras", []):
entry["clip_weight"] = clip_weights.get(str(entry.get("name", "")), 1.0)
print(f"[FunPackStudio] {lora_apply_status}")
except Exception as e:
print(f"[FunPackStudio] Apply LoRA weights failed: {e}")
loader = self._get_lora_loader()
if loader is not None and active_lora_stack is not None:
try:
model, clip, active_lora_stack, lora_load_status = loader.load_loras(
model, active_lora_stack, clip=clip
)
print(f"[FunPackStudio] {lora_load_status}")
except Exception as e:
print(f"[FunPackStudio] LoRA load failed: {e}")
active_lora_stack = raw_stack
# --- Refiner V2 ---
refiner = FunPackVideoRefinerV2()
@@ -12054,4 +12088,4 @@ class FunPackStudio:
except Exception as e:
status = f"{status}\nConditioning adjust failed: {e}"
return (cond, status, training_info, loss_graph, encoded_prompts, out_model, seed, active_lora_stack)
return (cond, status, training_info, loss_graph, encoded_prompts, out_model, seed)
web/funpack_studio.js
+11 -2
View File
@@ -64,6 +64,7 @@ function defaultSettings() {
refiner: { mode: "Refine", advisor_mode: "Off", advisor_thinking: true, prompt_repair: true, im_feeling_lucky: false, reset_session: false, feedback_prompt: "", user_intent_prompt_override: "" },
advisor_llm: { enabled: false, model_path: "huihui-ai/Huihui-Qwen3-8B-abliterated-v2", dtype: "bfloat16" },
loras: [],
loras_config: { mode: "ltx2", per_block: false },
};
}
@@ -507,9 +508,17 @@ function openPanel(node) {
// LORA ─────────────────────────────────────────────────────────────────────
function renderLora() {
body.append(sectionTitle("LoRA List"));
body.append(sectionTitle("LoRA Settings"));
if (!settings.loras_config) settings.loras_config = { mode: "ltx2", per_block: false };
const modeSelect = selectEl(["ltx2", "wan"], settings.loras_config.mode);
modeSelect.addEventListener("change", () => { settings.loras_config.mode = modeSelect.value; });
body.append(row("Model type", modeSelect));
const perBlockToggle = toggleEl(settings.loras_config.per_block, "Per-block application");
perBlockToggle.inp.addEventListener("change", () => { settings.loras_config.per_block = perBlockToggle.inp.checked; });
body.append(row("Per-block", perBlockToggle.wrap));
body.append(el("div", "funpack-studio-hint",
"LoRAs are loaded in order each run. An external lora_stack input overrides this list entirely."));
"Studio applies LoRAs internally: session weight suggestions are read first, then LoRAs are loaded, then the direction patch is applied on top. An external lora_stack input bypasses this entirely."));
body.append(sectionTitle("LoRA List"));
const list = el("div", "funpack-studio-lora-list");