final changelog + changes to denoise wording

workflows/TXT2IMG/v4.0/TXT2IMG-ADetailer-v4.0-vslinx.json

@@ -3068,36 +3068,6 @@
       "color": "#fff0",
       "bgcolor": "#fff0"
     },
-    {
-      "id": 354,
-      "type": "MarkdownNote",
-      "pos": [
-        -1150,
-        970
-      ],
-      "size": [
-        370,
-        740
-      ],
-      "flags": {
-        "collapsed": false,
-        "pinned": true
-      },
-      "order": 44,
-      "mode": 0,
-      "inputs": [],
-      "outputs": [],
-      "title": "Improve further",
-      "properties": {
-        "widget_ue_connectable": {}
-      },
-      "widgets_values": [
-        "## Detailer Tips & Usage Guide\n\n### Prompt Behavior\n- If **body-part specific prompts** are **disabled**, the detailer uses the **standard positive/negative prompts**.\n- Detectors and checkpoints usually infer the content well through denoising.\n- However, **writing specific prompts** for body parts results in **better and more accurate** generations.\n- To enhance further, use a **Detailer LoRA** (see recommended ones to the left).\n- If no specific LoRA exists, use a **general Detailer** (EPS or V-Pred) and prompt what you want shown.\n\n> By default, **negative prompts** are shared with the main image generation.  \n> To use a custom one, connect a new **Text Prompt Node** in the  \n> `\"Backend - Detailer Preparation\"` group.\n\n---\n\n### Debug Preview\n\n- After generation, check the **debug area** for one node per detected body part.\n- Click the rectangle next to a node's name to view the **mask**.\n- If the preview:\n  - shows the **full image** → the detailer was **inactive**.\n  - is **black** → the model **found no match**.\n\n> If no match is found, try changing the detection model  \n> or tweaking the **Detailer parameters** below.\n\n---\n\n## Key Detailer Parameters\n\n- **`bbox_crop_factor`**: Controls how much context around the area is used.\n  - `1.0` = tight crop, `3.0` = broader context.\n  - Too low = lacks context; too high = loses detail.\n  - Optimal range: **1.0–3.0**.\n\n- **`denoise`**: Determines how much of the original is changed.\n  - `0.5` = balanced, `1.0` = full replacement.\n  - Mid-range is ideal to retain form while enhancing detail.\n\n- **Positive Prompt Node**:  \n  Use optional body-part positive prompts in the \"Detailer prompts\"-Group.  \n  Use **body-part specific LoRA’s** for best results.\n\n- **Matching Parts**:  \n  Use a **detection model** that recognized a pair of eyes, nipples, etc., to ensure consistency.\n\n- **Preserve Original Details**:\n  - Lower **CFG** or **denoise** to retain color/shape (e.g., iris).\n"
-      ],
-      "color": "#222",
-      "bgcolor": "#000",
-      "shape": 1
-    },
     {
       "id": 360,
       "type": "MarkdownNote",
@@ -3112,7 +3082,7 @@
       "flags": {
         "pinned": true
       },
-      "order": 45,
+      "order": 44,
       "mode": 0,
       "inputs": [],
       "outputs": [],
@@ -3332,7 +3302,7 @@
         102
       ],
       "flags": {},
-      "order": 46,
+      "order": 45,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -3379,7 +3349,7 @@
       "flags": {
         "collapsed": true
       },
-      "order": 47,
+      "order": 46,
       "mode": 4,
       "inputs": [
         {
@@ -3423,7 +3393,7 @@
       "flags": {
         "collapsed": true
       },
-      "order": 48,
+      "order": 47,
       "mode": 4,
       "inputs": [
         {
@@ -3466,7 +3436,7 @@
         90
       ],
       "flags": {},
-      "order": 49,
+      "order": 48,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -3569,7 +3539,7 @@
         30
       ],
       "flags": {},
-      "order": 50,
+      "order": 49,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -3624,7 +3594,7 @@
         90
       ],
       "flags": {},
-      "order": 51,
+      "order": 50,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -3678,7 +3648,7 @@
         30
       ],
       "flags": {},
-      "order": 52,
+      "order": 51,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -3735,7 +3705,7 @@
       "flags": {
         "collapsed": true
       },
-      "order": 53,
+      "order": 52,
       "mode": 0,
       "inputs": [
         {
@@ -3782,7 +3752,7 @@
       "flags": {
         "collapsed": true
       },
-      "order": 54,
+      "order": 53,
       "mode": 0,
       "inputs": [
         {
@@ -4190,7 +4160,7 @@
         30
       ],
       "flags": {},
-      "order": 55,
+      "order": 54,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -4439,7 +4409,7 @@
         30
       ],
       "flags": {},
-      "order": 56,
+      "order": 55,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -4485,7 +4455,7 @@
         30
       ],
       "flags": {},
-      "order": 57,
+      "order": 56,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -4734,7 +4704,7 @@
         30
       ],
       "flags": {},
-      "order": 58,
+      "order": 57,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -5389,7 +5359,7 @@
         30
       ],
       "flags": {},
-      "order": 59,
+      "order": 58,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -5638,7 +5608,7 @@
         30
       ],
       "flags": {},
-      "order": 60,
+      "order": 59,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -5887,7 +5857,7 @@
         30
       ],
       "flags": {},
-      "order": 61,
+      "order": 60,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -6247,7 +6217,7 @@
         170
       ],
       "flags": {},
-      "order": 62,
+      "order": 61,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -6524,7 +6494,7 @@
         110
       ],
       "flags": {},
-      "order": 63,
+      "order": 62,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -6564,7 +6534,7 @@
         30
       ],
       "flags": {},
-      "order": 64,
+      "order": 63,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -6610,7 +6580,7 @@
         210
       ],
       "flags": {},
-      "order": 65,
+      "order": 64,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -7089,7 +7059,7 @@
         30
       ],
       "flags": {},
-      "order": 66,
+      "order": 65,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -7135,7 +7105,7 @@
         30
       ],
       "flags": {},
-      "order": 67,
+      "order": 66,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -7215,7 +7185,7 @@
         250
       ],
       "flags": {},
-      "order": 68,
+      "order": 67,
       "mode": 0,
       "inputs": [],
       "outputs": [],
@@ -8148,7 +8118,7 @@
         350
       ],
       "flags": {},
-      "order": 69,
+      "order": 68,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -8194,7 +8164,7 @@
         350
       ],
       "flags": {},
-      "order": 70,
+      "order": 69,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -8238,7 +8208,7 @@
         350
       ],
       "flags": {},
-      "order": 71,
+      "order": 70,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -8285,7 +8255,7 @@
         350
       ],
       "flags": {},
-      "order": 72,
+      "order": 71,
       "mode": 0,
       "inputs": [],
       "outputs": [
@@ -8331,7 +8301,7 @@
       "flags": {
         "pinned": true
       },
-      "order": 73,
+      "order": 72,
       "mode": 0,
       "inputs": [],
       "outputs": [],
@@ -8345,6 +8315,36 @@
       "color": "#222",
       "bgcolor": "#000",
       "shape": 1
+    },
+    {
+      "id": 354,
+      "type": "MarkdownNote",
+      "pos": [
+        -1150,
+        970
+      ],
+      "size": [
+        370,
+        740
+      ],
+      "flags": {
+        "collapsed": false,
+        "pinned": true
+      },
+      "order": 73,
+      "mode": 0,
+      "inputs": [],
+      "outputs": [],
+      "title": "Improve further",
+      "properties": {
+        "widget_ue_connectable": {}
+      },
+      "widgets_values": [
+        "## Detailer Tips & Usage Guide\n\n### Prompt Behavior\n- If **body-part specific prompts** are **disabled**, the detailer uses the **standard positive/negative prompts**.\n- Detectors and checkpoints usually infer the content well through denoising.\n- However, **writing specific prompts** for body parts results in **better and more accurate** generations.\n- To enhance further, use a **Detailer LoRA** (see recommended ones to the left).\n- If no specific LoRA exists, use a **general Detailer** (EPS or V-Pred) and prompt what you want shown.\n\n> By default, **negative prompts** are shared with the main image generation.  \n> To use a custom one, connect a new **Text Prompt Node** in the  \n> `\"Backend - Detailer Preparation\"` group.\n\n---\n\n### Debug Preview\n\n- After generation, check the **debug area** for one node per detected body part.\n- Click the rectangle next to a node's name to view the **mask**.\n- If the preview:\n  - shows the **full image** → the detailer was **inactive**.\n  - is **black** → the model **found no match**.\n\n> If no match is found, try changing the detection model  \n> or tweaking the **Detailer parameters** below.\n\n---\n\n## Key Detailer Parameters\n\n- **`bbox_crop_factor`**: Controls how much context around the area is used.\n  - `1.0` = tight crop, `3.0` = broader context.\n  - Too low = lacks context; too high = loses detail.\n  - Optimal range: **1.0–3.0**.\n\n- **`denoise`**: Determines how much of the original is changed.\n  - `0.5` = balanced, `1.0` = full replacement.\n  - Low to mid-range is ideal to retain original form while enhancing detail.\n\n- **Positive Prompt Node**:  \n  Use optional body-part positive prompts in the \"Detailer prompts\"-Group.  \n  Use **body-part specific LoRA’s** for best results.\n\n- **Matching Parts**:  \n  Use a **detection model** that recognized a pair of eyes, nipples, etc., to ensure consistency.\n\n- **Preserve Original Details**:\n  - Lower **CFG** or **denoise** to retain color/shape (e.g., iris).\n"
+      ],
+      "color": "#222",
+      "bgcolor": "#000",
+      "shape": 1
     }
   ],
   "links": [
@@ -10648,8 +10648,8 @@
     "ds": {
       "scale": 0.7513148009015777,
       "offset": [
-        794.2195754938466,
-        505.19739065464347
+        1058.289747743748,
+        422.07853984569476
       ]
     },
     "ue_links": [

workflows/TXT2IMG/v4.0/changelog.md

@@ -1,6 +1,10 @@
-- Implementation of IPAdapter (Optional)
-- Manual & Recommendations + Requirements overhaul
-- Reorganization of Backend Logic for better state handling
-- Compact Group Nodes for ControlNet & IPAdapter Models
-- Fixed clip inheritance amongst optional Detailer LoRA's & Optional Detailer prompts
+- Implementation of IPAdapter for Style & Concept (Optional)
+- CFG, Sampler, Scheduler & Steps now directly affect detailer settings as well
+- Body parts now have their own denoise slider for better detailing
+    - low denoise value = keep form, inpaint area according to original
+    - mid-high denoise value = overwrite form, inpaint area according to prompt while ignoring original content
+- Manual & Recommendations + Requirements overhaul
+- Reorganization of Backend Logic for better state handling
+- Compact Group Nodes for ControlNet & IPAdapter Models
+- Fixed clip inheritance amongst optional Detailer LoRA's & Optional Detailer prompts
 - Debug section now shows full prompt preview at all times