Merge branch 'master' into add-coderabbit-config

Amp-Thread-ID: https://ampcode.com/threads/T-019c7915-2abf-743c-9c74-b93d87d63055

.coderabbit.yaml

@@ -1,7 +1,6 @@
 # yaml-language-server: $schema=https://coderabbit.ai/integrations/schema.v2.json
 language: "en-US"
 early_access: false
-tone_instructions: "Only comment on issues introduced by this PR's changes. Do not flag pre-existing problems in moved, re-indented, or reformatted code."
 
 reviews:
   profile: "chill"
@@ -36,14 +35,6 @@ reviews:
     - "!**/*.bat"
 
   path_instructions:
-    - path: "**"
-      instructions: |
-        IMPORTANT: Only comment on issues directly introduced by this PR's code changes.
-        Do NOT flag pre-existing issues in code that was merely moved, re-indented,
-        de-indented, or reformatted without logic changes. If code appears in the diff
-        only due to whitespace or structural reformatting (e.g., removing a `with:` block),
-        treat it as unchanged. Contributors should not feel obligated to address
-        pre-existing issues outside the scope of their contribution.
     - path: "comfy/**"
       instructions: |
         Core ML/diffusion engine. Focus on:
@@ -83,11 +74,7 @@ reviews:
   auto_review:
     enabled: true
     auto_incremental_review: true
-    drafts: false
-    ignore_title_keywords:
-      - "WIP"
-      - "DO NOT REVIEW"
-      - "DO NOT MERGE"
+    drafts: true
 
   finishing_touches:
     docstrings:
@@ -97,7 +84,7 @@ reviews:
 
   tools:
     ruff:
-      enabled: false
+      enabled: true
     pylint:
       enabled: false
     flake8:

README.md

@@ -229,9 +229,9 @@ AMD users can install rocm and pytorch with pip if you don't have it already ins
 
 ```pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm7.1```
 
-This is the command to install the nightly with ROCm 7.2 which might have some performance improvements:
+This is the command to install the nightly with ROCm 7.1 which might have some performance improvements:
 
-```pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm7.2```
+```pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm7.1```
 
 
 ### AMD GPUs (Experimental: Windows and Linux), RDNA 3, 3.5 and 4 only.

app/subgraph_manager.py

@@ -53,7 +53,7 @@ class SubgraphManager:
         return entry_id, entry
 
     async def load_entry_data(self, entry: SubgraphEntry):
-        with open(entry['path'], 'r', encoding='utf-8') as f:
+        with open(entry['path'], 'r') as f:
             entry['data'] = f.read()
         return entry
 

blueprints/Image Channels.json

@@ -1 +1 @@
-{"revision": 0, "last_node_id": 29, "last_link_id": 0, "nodes": [{"id": 29, "type": "4c9d6ea4-b912-40e5-8766-6793a9758c53", "pos": [1970, -230], "size": [180, 86], "flags": {}, "order": 5, "mode": 0, "inputs": [{"label": "image", "localized_name": "images.image0", "name": "images.image0", "type": "IMAGE", "link": null}], "outputs": [{"label": "R", "localized_name": "IMAGE0", "name": "IMAGE0", "type": "IMAGE", "links": []}, {"label": "G", "localized_name": "IMAGE1", "name": "IMAGE1", "type": "IMAGE", "links": []}, {"label": "B", "localized_name": "IMAGE2", "name": "IMAGE2", "type": "IMAGE", "links": []}, {"label": "A", "localized_name": "IMAGE3", "name": "IMAGE3", "type": "IMAGE", "links": []}], "title": "Image Channels", "properties": {"proxyWidgets": []}, "widgets_values": []}], "links": [], "version": 0.4, "definitions": {"subgraphs": [{"id": "4c9d6ea4-b912-40e5-8766-6793a9758c53", "version": 1, "state": {"lastGroupId": 0, "lastNodeId": 28, "lastLinkId": 39, "lastRerouteId": 0}, "revision": 0, "config": {}, "name": "Image Channels", "inputNode": {"id": -10, "bounding": [1820, -185, 120, 60]}, "outputNode": {"id": -20, "bounding": [2460, -215, 120, 120]}, "inputs": [{"id": "3522932b-2d86-4a1f-a02a-cb29f3a9d7fe", "name": "images.image0", "type": "IMAGE", "linkIds": [39], "localized_name": "images.image0", "label": "image", "pos": [1920, -165]}], "outputs": [{"id": "605cb9c3-b065-4d9b-81d2-3ec331889b2b", "name": "IMAGE0", "type": "IMAGE", "linkIds": [26], "localized_name": "IMAGE0", "label": "R", "pos": [2480, -195]}, {"id": "fb44a77e-0522-43e9-9527-82e7465b3596", "name": "IMAGE1", "type": "IMAGE", "linkIds": [27], "localized_name": "IMAGE1", "label": "G", "pos": [2480, -175]}, {"id": "81460ee6-0131-402a-874f-6bf3001fc4ff", "name": "IMAGE2", "type": "IMAGE", "linkIds": [28], "localized_name": "IMAGE2", "label": "B", "pos": [2480, -155]}, {"id": "ae690246-80d4-4951-b1d9-9306d8a77417", "name": "IMAGE3", "type": "IMAGE", "linkIds": [29], "localized_name": "IMAGE3", "label": "A", "pos": [2480, -135]}], "widgets": [], "nodes": [{"id": 23, "type": "GLSLShader", "pos": [2000, -330], "size": [400, 172], "flags": {}, "order": 0, "mode": 0, "inputs": [{"label": "image", "localized_name": "images.image0", "name": "images.image0", "type": "IMAGE", "link": 39}, {"localized_name": "fragment_shader", "name": "fragment_shader", "type": "STRING", "widget": {"name": "fragment_shader"}, "link": null}, {"localized_name": "size_mode", "name": "size_mode", "type": "COMFY_DYNAMICCOMBO_V3", "widget": {"name": "size_mode"}, "link": null}, {"label": "image1", "localized_name": "images.image1", "name": "images.image1", "shape": 7, "type": "IMAGE", "link": null}], "outputs": [{"label": "R", "localized_name": "IMAGE0", "name": "IMAGE0", "type": "IMAGE", "links": [26]}, {"label": "G", "localized_name": "IMAGE1", "name": "IMAGE1", "type": "IMAGE", "links": [27]}, {"label": "B", "localized_name": "IMAGE2", "name": "IMAGE2", "type": "IMAGE", "links": [28]}, {"label": "A", "localized_name": "IMAGE3", "name": "IMAGE3", "type": "IMAGE", "links": [29]}], "properties": {"Node name for S&R": "GLSLShader"}, "widgets_values": ["#version 300 es\nprecision highp float;\n\nuniform sampler2D u_image0;\n\nin vec2 v_texCoord;\nlayout(location = 0) out vec4 fragColor0;\nlayout(location = 1) out vec4 fragColor1;\nlayout(location = 2) out vec4 fragColor2;\nlayout(location = 3) out vec4 fragColor3;\n\nvoid main() {\n  vec4 color = texture(u_image0, v_texCoord);\n  // Output each channel as grayscale to separate render targets\n  fragColor0 = vec4(vec3(color.r), 1.0);  // Red channel\n  fragColor1 = vec4(vec3(color.g), 1.0);  // Green channel\n  fragColor2 = vec4(vec3(color.b), 1.0);  // Blue channel\n  fragColor3 = vec4(vec3(color.a), 1.0);  // Alpha channel\n}\n", "from_input"]}], "groups": [], "links": [{"id": 39, "origin_id": -10, "origin_slot": 0, "target_id": 23, "target_slot": 0, "type": "IMAGE"}, {"id": 26, "origin_id": 23, "origin_slot": 0, "target_id": -20, "target_slot": 0, "type": "IMAGE"}, {"id": 27, "origin_id": 23, "origin_slot": 1, "target_id": -20, "target_slot": 1, "type": "IMAGE"}, {"id": 28, "origin_id": 23, "origin_slot": 2, "target_id": -20, "target_slot": 2, "type": "IMAGE"}, {"id": 29, "origin_id": 23, "origin_slot": 3, "target_id": -20, "target_slot": 3, "type": "IMAGE"}], "extra": {"workflowRendererVersion": "LG"}, "category": "Image Tools/Color adjust"}]}}
+{"revision": 0, "last_node_id": 29, "last_link_id": 0, "nodes": [{"id": 29, "type": "4c9d6ea4-b912-40e5-8766-6793a9758c53", "pos": [1970, -230], "size": [180, 86], "flags": {}, "order": 5, "mode": 0, "inputs": [{"label": "image", "localized_name": "images.image0", "name": "images.image0", "type": "IMAGE", "link": null}], "outputs": [{"label": "R", "localized_name": "IMAGE0", "name": "IMAGE0", "type": "IMAGE", "links": []}, {"label": "G", "localized_name": "IMAGE1", "name": "IMAGE1", "type": "IMAGE", "links": []}, {"label": "B", "localized_name": "IMAGE2", "name": "IMAGE2", "type": "IMAGE", "links": []}, {"label": "A", "localized_name": "IMAGE3", "name": "IMAGE3", "type": "IMAGE", "links": []}], "title": "Image Channels", "properties": {"proxyWidgets": []}, "widgets_values": []}], "links": [], "version": 0.4, "definitions": {"subgraphs": [{"id": "4c9d6ea4-b912-40e5-8766-6793a9758c53", "version": 1, "state": {"lastGroupId": 0, "lastNodeId": 28, "lastLinkId": 39, "lastRerouteId": 0}, "revision": 0, "config": {}, "name": "Image Channels", "inputNode": {"id": -10, "bounding": [1820, -185, 120, 60]}, "outputNode": {"id": -20, "bounding": [2460, -215, 120, 120]}, "inputs": [{"id": "3522932b-2d86-4a1f-a02a-cb29f3a9d7fe", "name": "images.image0", "type": "IMAGE", "linkIds": [39], "localized_name": "images.image0", "label": "image", "pos": [1920, -165]}], "outputs": [{"id": "605cb9c3-b065-4d9b-81d2-3ec331889b2b", "name": "IMAGE0", "type": "IMAGE", "linkIds": [26], "localized_name": "IMAGE0", "label": "R", "pos": [2480, -195]}, {"id": "fb44a77e-0522-43e9-9527-82e7465b3596", "name": "IMAGE1", "type": "IMAGE", "linkIds": [27], "localized_name": "IMAGE1", "label": "G", "pos": [2480, -175]}, {"id": "81460ee6-0131-402a-874f-6bf3001fc4ff", "name": "IMAGE2", "type": "IMAGE", "linkIds": [28], "localized_name": "IMAGE2", "label": "B", "pos": [2480, -155]}, {"id": "ae690246-80d4-4951-b1d9-9306d8a77417", "name": "IMAGE3", "type": "IMAGE", "linkIds": [29], "localized_name": "IMAGE3", "label": "A", "pos": [2480, -135]}], "widgets": [], "nodes": [{"id": 23, "type": "GLSLShader", "pos": [2000, -330], "size": [400, 172], "flags": {}, "order": 0, "mode": 0, "inputs": [{"label": "image", "localized_name": "images.image0", "name": "images.image0", "type": "IMAGE", "link": 39}, {"localized_name": "fragment_shader", "name": "fragment_shader", "type": "STRING", "widget": {"name": "fragment_shader"}, "link": null}, {"localized_name": "size_mode", "name": "size_mode", "type": "COMFY_DYNAMICCOMBO_V3", "widget": {"name": "size_mode"}, "link": null}, {"label": "image1", "localized_name": "images.image1", "name": "images.image1", "shape": 7, "type": "IMAGE", "link": null}], "outputs": [{"label": "R", "localized_name": "IMAGE0", "name": "IMAGE0", "type": "IMAGE", "links": [26]}, {"label": "G", "localized_name": "IMAGE1", "name": "IMAGE1", "type": "IMAGE", "links": [27]}, {"label": "B", "localized_name": "IMAGE2", "name": "IMAGE2", "type": "IMAGE", "links": [28]}, {"label": "A", "localized_name": "IMAGE3", "name": "IMAGE3", "type": "IMAGE", "links": [29]}], "properties": {"Node name for S&R": "GLSLShader"}, "widgets_values": ["#version 300 es\nprecision highp float;\n\nuniform sampler2D u_image0;\n\nin vec2 v_texCoord;\nlayout(location = 0) out vec4 fragColor0;\nlayout(location = 1) out vec4 fragColor1;\nlayout(location = 2) out vec4 fragColor2;\nlayout(location = 3) out vec4 fragColor3;\n\nvoid main() {\n  vec4 color = texture(u_image0, v_texCoord);\n  // Output each channel as grayscale to separate render targets\n  fragColor0 = vec4(vec3(color.r), 1.0);  // Red channel\n  fragColor1 = vec4(vec3(color.g), 1.0);  // Green channel\n  fragColor2 = vec4(vec3(color.b), 1.0);  // Blue channel\n  fragColor3 = vec4(vec3(color.a), 1.0);  // Alpha channel\n}\n", "from_input"]}], "groups": [], "links": [{"id": 39, "origin_id": -10, "origin_slot": 0, "target_id": 23, "target_slot": 0, "type": "IMAGE"}, {"id": 26, "origin_id": 23, "origin_slot": 0, "target_id": -20, "target_slot": 0, "type": "IMAGE"}, {"id": 27, "origin_id": 23, "origin_slot": 1, "target_id": -20, "target_slot": 1, "type": "IMAGE"}, {"id": 28, "origin_id": 23, "origin_slot": 2, "target_id": -20, "target_slot": 2, "type": "IMAGE"}, {"id": 29, "origin_id": 23, "origin_slot": 3, "target_id": -20, "target_slot": 3, "type": "IMAGE"}], "extra": {"workflowRendererVersion": "LG"}}]}}

blueprints/Prompt Enhance.json

@@ -1 +0,0 @@
-{"revision": 0, "last_node_id": 15, "last_link_id": 0, "nodes": [{"id": 15, "type": "24d8bbfd-39d4-4774-bff0-3de40cc7a471", "pos": [-1490, 2040], "size": [400, 260], "flags": {}, "order": 0, "mode": 0, "inputs": [{"name": "prompt", "type": "STRING", "widget": {"name": "prompt"}, "link": null}, {"label": "reference images", "name": "images", "type": "IMAGE", "link": null}], "outputs": [{"name": "STRING", "type": "STRING", "links": null}], "title": "Prompt Enhance", "properties": {"proxyWidgets": [["-1", "prompt"]], "cnr_id": "comfy-core", "ver": "0.14.1"}, "widgets_values": [""]}], "links": [], "version": 0.4, "definitions": {"subgraphs": [{"id": "24d8bbfd-39d4-4774-bff0-3de40cc7a471", "version": 1, "state": {"lastGroupId": 0, "lastNodeId": 15, "lastLinkId": 14, "lastRerouteId": 0}, "revision": 0, "config": {}, "name": "Prompt Enhance", "inputNode": {"id": -10, "bounding": [-2170, 2110, 138.876953125, 80]}, "outputNode": {"id": -20, "bounding": [-640, 2110, 120, 60]}, "inputs": [{"id": "aeab7216-00e0-4528-a09b-bba50845c5a6", "name": "prompt", "type": "STRING", "linkIds": [11], "pos": [-2051.123046875, 2130]}, {"id": "7b73fd36-aa31-4771-9066-f6c83879994b", "name": "images", "type": "IMAGE", "linkIds": [14], "label": "reference images", "pos": [-2051.123046875, 2150]}], "outputs": [{"id": "c7b0d930-68a1-48d1-b496-0519e5837064", "name": "STRING", "type": "STRING", "linkIds": [13], "pos": [-620, 2130]}], "widgets": [], "nodes": [{"id": 11, "type": "GeminiNode", "pos": [-1560, 1990], "size": [470, 470], "flags": {}, "order": 0, "mode": 0, "inputs": [{"localized_name": "images", "name": "images", "shape": 7, "type": "IMAGE", "link": 14}, {"localized_name": "audio", "name": "audio", "shape": 7, "type": "AUDIO", "link": null}, {"localized_name": "video", "name": "video", "shape": 7, "type": "VIDEO", "link": null}, {"localized_name": "files", "name": "files", "shape": 7, "type": "GEMINI_INPUT_FILES", "link": null}, {"localized_name": "prompt", "name": "prompt", "type": "STRING", "widget": {"name": "prompt"}, "link": 11}, {"localized_name": "model", "name": "model", "type": "COMBO", "widget": {"name": "model"}, "link": null}, {"localized_name": "seed", "name": "seed", "type": "INT", "widget": {"name": "seed"}, "link": null}, {"localized_name": "system_prompt", "name": "system_prompt", "shape": 7, "type": "STRING", "widget": {"name": "system_prompt"}, "link": null}], "outputs": [{"localized_name": "STRING", "name": "STRING", "type": "STRING", "links": [13]}], "properties": {"cnr_id": "comfy-core", "ver": "0.14.1", "Node name for S&R": "GeminiNode"}, "widgets_values": ["", "gemini-3-pro-preview", 42, "randomize", "You are an expert in prompt writing.\nBased on the input, rewrite the user's input into a detailed prompt.\nincluding camera settings, lighting, composition, and style.\nReturn the prompt only"], "color": "#432", "bgcolor": "#653"}], "groups": [], "links": [{"id": 11, "origin_id": -10, "origin_slot": 0, "target_id": 11, "target_slot": 4, "type": "STRING"}, {"id": 13, "origin_id": 11, "origin_slot": 0, "target_id": -20, "target_slot": 0, "type": "STRING"}, {"id": 14, "origin_id": -10, "origin_slot": 1, "target_id": 11, "target_slot": 0, "type": "IMAGE"}], "extra": {"workflowRendererVersion": "LG"}, "category": "Text generation/Prompt enhance"}]}, "extra": {}}

blueprints/Sharpen.json

@@ -1 +1 @@
-{"revision": 0, "last_node_id": 25, "last_link_id": 0, "nodes": [{"id": 25, "type": "621ba4e2-22a8-482d-a369-023753198b7b", "pos": [4610, -790], "size": [230, 58], "flags": {}, "order": 4, "mode": 0, "inputs": [{"label": "image", "localized_name": "images.image0", "name": "images.image0", "type": "IMAGE", "link": null}], "outputs": [{"label": "IMAGE", "localized_name": "IMAGE0", "name": "IMAGE0", "type": "IMAGE", "links": []}], "title": "Sharpen", "properties": {"proxyWidgets": [["24", "value"]]}, "widgets_values": []}], "links": [], "version": 0.4, "definitions": {"subgraphs": [{"id": "621ba4e2-22a8-482d-a369-023753198b7b", "version": 1, "state": {"lastGroupId": 0, "lastNodeId": 24, "lastLinkId": 36, "lastRerouteId": 0}, "revision": 0, "config": {}, "name": "Sharpen", "inputNode": {"id": -10, "bounding": [4090, -825, 120, 60]}, "outputNode": {"id": -20, "bounding": [5150, -825, 120, 60]}, "inputs": [{"id": "37011fb7-14b7-4e0e-b1a0-6a02e8da1fd7", "name": "images.image0", "type": "IMAGE", "linkIds": [34], "localized_name": "images.image0", "label": "image", "pos": [4190, -805]}], "outputs": [{"id": "e9182b3f-635c-4cd4-a152-4b4be17ae4b9", "name": "IMAGE0", "type": "IMAGE", "linkIds": [35], "localized_name": "IMAGE0", "label": "IMAGE", "pos": [5170, -805]}], "widgets": [], "nodes": [{"id": 24, "type": "PrimitiveFloat", "pos": [4280, -1240], "size": [270, 58], "flags": {}, "order": 0, "mode": 0, "inputs": [{"label": "strength", "localized_name": "value", "name": "value", "type": "FLOAT", "widget": {"name": "value"}, "link": null}], "outputs": [{"localized_name": "FLOAT", "name": "FLOAT", "type": "FLOAT", "links": [36]}], "properties": {"Node name for S&R": "PrimitiveFloat", "min": 0, "max": 3, "precision": 2, "step": 0.05}, "widgets_values": [0.5]}, {"id": 23, "type": "GLSLShader", "pos": [4570, -1240], "size": [370, 192], "flags": {}, "order": 1, "mode": 0, "inputs": [{"label": "image0", "localized_name": "images.image0", "name": "images.image0", "type": "IMAGE", "link": 34}, {"label": "image1", "localized_name": "images.image1", "name": "images.image1", "shape": 7, "type": "IMAGE", "link": null}, {"label": "u_float0", "localized_name": "floats.u_float0", "name": "floats.u_float0", "shape": 7, "type": "FLOAT", "link": 36}, {"label": "u_float1", "localized_name": "floats.u_float1", "name": "floats.u_float1", "shape": 7, "type": "FLOAT", "link": null}, {"label": "u_int0", "localized_name": "ints.u_int0", "name": "ints.u_int0", "shape": 7, "type": "INT", "link": null}, {"localized_name": "fragment_shader", "name": "fragment_shader", "type": "STRING", "widget": {"name": "fragment_shader"}, "link": null}, {"localized_name": "size_mode", "name": "size_mode", "type": "COMFY_DYNAMICCOMBO_V3", "widget": {"name": "size_mode"}, "link": null}], "outputs": [{"localized_name": "IMAGE0", "name": "IMAGE0", "type": "IMAGE", "links": [35]}, {"localized_name": "IMAGE1", "name": "IMAGE1", "type": "IMAGE", "links": null}, {"localized_name": "IMAGE2", "name": "IMAGE2", "type": "IMAGE", "links": null}, {"localized_name": "IMAGE3", "name": "IMAGE3", "type": "IMAGE", "links": null}], "properties": {"Node name for S&R": "GLSLShader"}, "widgets_values": ["#version 300 es\nprecision highp float;\n\nuniform sampler2D u_image0;\nuniform vec2 u_resolution;\nuniform float u_float0;  // strength [0.0 – 2.0] typical: 0.3–1.0\n\nin vec2 v_texCoord;\nlayout(location = 0) out vec4 fragColor0;\n\nvoid main() {\n    vec2 texel = 1.0 / u_resolution;\n    \n    // Sample center and neighbors\n    vec4 center = texture(u_image0, v_texCoord);\n    vec4 top    = texture(u_image0, v_texCoord + vec2( 0.0, -texel.y));\n    vec4 bottom = texture(u_image0, v_texCoord + vec2( 0.0,  texel.y));\n    vec4 left   = texture(u_image0, v_texCoord + vec2(-texel.x,  0.0));\n    vec4 right  = texture(u_image0, v_texCoord + vec2( texel.x,  0.0));\n    \n    // Edge enhancement (Laplacian)\n    vec4 edges = center * 4.0 - top - bottom - left - right;\n    \n    // Add edges back scaled by strength\n    vec4 sharpened = center + edges * u_float0;\n    \n    fragColor0 = vec4(clamp(sharpened.rgb, 0.0, 1.0), center.a);\n}", "from_input"]}], "groups": [], "links": [{"id": 36, "origin_id": 24, "origin_slot": 0, "target_id": 23, "target_slot": 2, "type": "FLOAT"}, {"id": 34, "origin_id": -10, "origin_slot": 0, "target_id": 23, "target_slot": 0, "type": "IMAGE"}, {"id": 35, "origin_id": 23, "origin_slot": 0, "target_id": -20, "target_slot": 0, "type": "IMAGE"}], "extra": {"workflowRendererVersion": "LG"}, "category": "Image Tools/Sharpen"}]}}
+{"revision":0,"last_node_id":25,"last_link_id":0,"nodes":[{"id":25,"type":"621ba4e2-22a8-482d-a369-023753198b7b","pos":[4610,-790],"size":[230,58],"flags":{},"order":4,"mode":0,"inputs":[{"label":"image","localized_name":"images.image0","name":"images.image0","type":"IMAGE","link":null}],"outputs":[{"label":"IMAGE","localized_name":"IMAGE0","name":"IMAGE0","type":"IMAGE","links":[]}],"title":"Sharpen","properties":{"proxyWidgets":[["24","value"]]},"widgets_values":[]}],"links":[],"version":0.4,"definitions":{"subgraphs":[{"id":"621ba4e2-22a8-482d-a369-023753198b7b","version":1,"state":{"lastGroupId":0,"lastNodeId":24,"lastLinkId":36,"lastRerouteId":0},"revision":0,"config":{},"name":"Sharpen","inputNode":{"id":-10,"bounding":[4090,-825,120,60]},"outputNode":{"id":-20,"bounding":[5150,-825,120,60]},"inputs":[{"id":"37011fb7-14b7-4e0e-b1a0-6a02e8da1fd7","name":"images.image0","type":"IMAGE","linkIds":[34],"localized_name":"images.image0","label":"image","pos":[4190,-805]}],"outputs":[{"id":"e9182b3f-635c-4cd4-a152-4b4be17ae4b9","name":"IMAGE0","type":"IMAGE","linkIds":[35],"localized_name":"IMAGE0","label":"IMAGE","pos":[5170,-805]}],"widgets":[],"nodes":[{"id":24,"type":"PrimitiveFloat","pos":[4280,-1240],"size":[270,58],"flags":{},"order":0,"mode":0,"inputs":[{"label":"strength","localized_name":"value","name":"value","type":"FLOAT","widget":{"name":"value"},"link":null}],"outputs":[{"localized_name":"FLOAT","name":"FLOAT","type":"FLOAT","links":[36]}],"properties":{"Node name for S&R":"PrimitiveFloat","min":0,"max":3,"precision":2,"step":0.05},"widgets_values":[0.5]},{"id":23,"type":"GLSLShader","pos":[4570,-1240],"size":[370,192],"flags":{},"order":1,"mode":0,"inputs":[{"label":"image0","localized_name":"images.image0","name":"images.image0","type":"IMAGE","link":34},{"label":"image1","localized_name":"images.image1","name":"images.image1","shape":7,"type":"IMAGE","link":null},{"label":"u_float0","localized_name":"floats.u_float0","name":"floats.u_float0","shape":7,"type":"FLOAT","link":36},{"label":"u_float1","localized_name":"floats.u_float1","name":"floats.u_float1","shape":7,"type":"FLOAT","link":null},{"label":"u_int0","localized_name":"ints.u_int0","name":"ints.u_int0","shape":7,"type":"INT","link":null},{"localized_name":"fragment_shader","name":"fragment_shader","type":"STRING","widget":{"name":"fragment_shader"},"link":null},{"localized_name":"size_mode","name":"size_mode","type":"COMFY_DYNAMICCOMBO_V3","widget":{"name":"size_mode"},"link":null}],"outputs":[{"localized_name":"IMAGE0","name":"IMAGE0","type":"IMAGE","links":[35]},{"localized_name":"IMAGE1","name":"IMAGE1","type":"IMAGE","links":null},{"localized_name":"IMAGE2","name":"IMAGE2","type":"IMAGE","links":null},{"localized_name":"IMAGE3","name":"IMAGE3","type":"IMAGE","links":null}],"properties":{"Node name for S&R":"GLSLShader"},"widgets_values":["#version 300 es\nprecision highp float;\n\nuniform sampler2D u_image0;\nuniform vec2 u_resolution;\nuniform float u_float0;  // strength [0.0 – 2.0] typical: 0.3–1.0\n\nin vec2 v_texCoord;\nlayout(location = 0) out vec4 fragColor0;\n\nvoid main() {\n    vec2 texel = 1.0 / u_resolution;\n    \n    // Sample center and neighbors\n    vec4 center = texture(u_image0, v_texCoord);\n    vec4 top    = texture(u_image0, v_texCoord + vec2( 0.0, -texel.y));\n    vec4 bottom = texture(u_image0, v_texCoord + vec2( 0.0,  texel.y));\n    vec4 left   = texture(u_image0, v_texCoord + vec2(-texel.x,  0.0));\n    vec4 right  = texture(u_image0, v_texCoord + vec2( texel.x,  0.0));\n    \n    // Edge enhancement (Laplacian)\n    vec4 edges = center * 4.0 - top - bottom - left - right;\n    \n    // Add edges back scaled by strength\n    vec4 sharpened = center + edges * u_float0;\n    \n    fragColor0 = vec4(clamp(sharpened.rgb, 0.0, 1.0), center.a);\n}","from_input"]}],"groups":[],"links":[{"id":36,"origin_id":24,"origin_slot":0,"target_id":23,"target_slot":2,"type":"FLOAT"},{"id":34,"origin_id":-10,"origin_slot":0,"target_id":23,"target_slot":0,"type":"IMAGE"},{"id":35,"origin_id":23,"origin_slot":0,"target_id":-20,"target_slot":0,"type":"IMAGE"}],"extra":{"workflowRendererVersion":"LG"}}]}}

blueprints/Video Upscale(GAN x4).json

@@ -1 +0,0 @@
-{"revision": 0, "last_node_id": 13, "last_link_id": 0, "nodes": [{"id": 13, "type": "cf95b747-3e17-46cb-8097-cac60ff9b2e1", "pos": [1120, 330], "size": [240, 58], "flags": {}, "order": 3, "mode": 0, "inputs": [{"localized_name": "video", "name": "video", "type": "VIDEO", "link": null}, {"name": "model_name", "type": "COMBO", "widget": {"name": "model_name"}, "link": null}], "outputs": [{"localized_name": "VIDEO", "name": "VIDEO", "type": "VIDEO", "links": []}], "title": "Video Upscale(GAN x4)", "properties": {"proxyWidgets": [["-1", "model_name"]], "cnr_id": "comfy-core", "ver": "0.14.1"}, "widgets_values": ["RealESRGAN_x4plus.safetensors"]}], "links": [], "version": 0.4, "definitions": {"subgraphs": [{"id": "cf95b747-3e17-46cb-8097-cac60ff9b2e1", "version": 1, "state": {"lastGroupId": 0, "lastNodeId": 13, "lastLinkId": 19, "lastRerouteId": 0}, "revision": 0, "config": {}, "name": "Video Upscale(GAN x4)", "inputNode": {"id": -10, "bounding": [550, 460, 120, 80]}, "outputNode": {"id": -20, "bounding": [1490, 460, 120, 60]}, "inputs": [{"id": "666d633e-93e7-42dc-8d11-2b7b99b0f2a6", "name": "video", "type": "VIDEO", "linkIds": [10], "localized_name": "video", "pos": [650, 480]}, {"id": "2e23a087-caa8-4d65-99e6-662761aa905a", "name": "model_name", "type": "COMBO", "linkIds": [19], "pos": [650, 500]}], "outputs": [{"id": "0c1768ea-3ec2-412f-9af6-8e0fa36dae70", "name": "VIDEO", "type": "VIDEO", "linkIds": [15], "localized_name": "VIDEO", "pos": [1510, 480]}], "widgets": [], "nodes": [{"id": 2, "type": "ImageUpscaleWithModel", "pos": [1110, 450], "size": [320, 46], "flags": {}, "order": 1, "mode": 0, "inputs": [{"localized_name": "upscale_model", "name": "upscale_model", "type": "UPSCALE_MODEL", "link": 1}, {"localized_name": "image", "name": "image", "type": "IMAGE", "link": 14}], "outputs": [{"localized_name": "IMAGE", "name": "IMAGE", "type": "IMAGE", "links": [13]}], "properties": {"cnr_id": "comfy-core", "ver": "0.10.0", "Node name for S&R": "ImageUpscaleWithModel"}}, {"id": 11, "type": "CreateVideo", "pos": [1110, 550], "size": [320, 78], "flags": {}, "order": 3, "mode": 0, "inputs": [{"localized_name": "images", "name": "images", "type": "IMAGE", "link": 13}, {"localized_name": "audio", "name": "audio", "shape": 7, "type": "AUDIO", "link": 16}, {"localized_name": "fps", "name": "fps", "type": "FLOAT", "widget": {"name": "fps"}, "link": 12}], "outputs": [{"localized_name": "VIDEO", "name": "VIDEO", "type": "VIDEO", "links": [15]}], "properties": {"cnr_id": "comfy-core", "ver": "0.10.0", "Node name for S&R": "CreateVideo"}, "widgets_values": [30]}, {"id": 10, "type": "GetVideoComponents", "pos": [1110, 330], "size": [320, 70], "flags": {}, "order": 2, "mode": 0, "inputs": [{"localized_name": "video", "name": "video", "type": "VIDEO", "link": 10}], "outputs": [{"localized_name": "images", "name": "images", "type": "IMAGE", "links": [14]}, {"localized_name": "audio", "name": "audio", "type": "AUDIO", "links": [16]}, {"localized_name": "fps", "name": "fps", "type": "FLOAT", "links": [12]}], "properties": {"cnr_id": "comfy-core", "ver": "0.10.0", "Node name for S&R": "GetVideoComponents"}}, {"id": 1, "type": "UpscaleModelLoader", "pos": [750, 450], "size": [280, 60], "flags": {}, "order": 0, "mode": 0, "inputs": [{"localized_name": "model_name", "name": "model_name", "type": "COMBO", "widget": {"name": "model_name"}, "link": 19}], "outputs": [{"localized_name": "UPSCALE_MODEL", "name": "UPSCALE_MODEL", "type": "UPSCALE_MODEL", "links": [1]}], "properties": {"cnr_id": "comfy-core", "ver": "0.10.0", "Node name for S&R": "UpscaleModelLoader", "models": [{"name": "RealESRGAN_x4plus.safetensors", "url": "https://huggingface.co/Comfy-Org/Real-ESRGAN_repackaged/resolve/main/RealESRGAN_x4plus.safetensors", "directory": "upscale_models"}]}, "widgets_values": ["RealESRGAN_x4plus.safetensors"]}], "groups": [], "links": [{"id": 1, "origin_id": 1, "origin_slot": 0, "target_id": 2, "target_slot": 0, "type": "UPSCALE_MODEL"}, {"id": 14, "origin_id": 10, "origin_slot": 0, "target_id": 2, "target_slot": 1, "type": "IMAGE"}, {"id": 13, "origin_id": 2, "origin_slot": 0, "target_id": 11, "target_slot": 0, "type": "IMAGE"}, {"id": 16, "origin_id": 10, "origin_slot": 1, "target_id": 11, "target_slot": 1, "type": "AUDIO"}, {"id": 12, "origin_id": 10, "origin_slot": 2, "target_id": 11, "target_slot": 2, "type": "FLOAT"}, {"id": 10, "origin_id": -10, "origin_slot": 0, "target_id": 10, "target_slot": 0, "type": "VIDEO"}, {"id": 15, "origin_id": 11, "origin_slot": 0, "target_id": -20, "target_slot": 0, "type": "VIDEO"}, {"id": 19, "origin_id": -10, "origin_slot": 1, "target_id": 1, "target_slot": 0, "type": "COMBO"}], "extra": {"workflowRendererVersion": "LG"}, "category": "Video generation and editing/Enhance video"}]}, "extra": {}}

comfy/comfy_types/node_typing.py

@@ -176,8 +176,6 @@ class InputTypeOptions(TypedDict):
     """COMBO type only. Specifies the configuration for a multi-select widget.
     Available after ComfyUI frontend v1.13.4
     https://github.com/Comfy-Org/ComfyUI_frontend/pull/2987"""
-    gradient_stops: NotRequired[list[list[float]]]
-    """Gradient color stops for gradientslider display mode. Each stop is [offset, r, g, b] (``FLOAT``)."""
 
 
 class HiddenInputTypeDict(TypedDict):

comfy/ldm/lightricks/av_model.py

@@ -9,7 +9,6 @@ from comfy.ldm.lightricks.model import (
     LTXVModel,
 )
 from comfy.ldm.lightricks.symmetric_patchifier import AudioPatchifier
-from comfy.ldm.lightricks.embeddings_connector import Embeddings1DConnector
 import comfy.ldm.common_dit
 
 class CompressedTimestep:
@@ -451,29 +450,6 @@ class LTXAVModel(LTXVModel):
             operations=self.operations,
         )
 
-        self.audio_embeddings_connector = Embeddings1DConnector(
-            split_rope=True,
-            double_precision_rope=True,
-            dtype=dtype,
-            device=device,
-            operations=self.operations,
-        )
-
-        self.video_embeddings_connector = Embeddings1DConnector(
-            split_rope=True,
-            double_precision_rope=True,
-            dtype=dtype,
-            device=device,
-            operations=self.operations,
-        )
-
-    def preprocess_text_embeds(self, context):
-        if context.shape[-1] == self.caption_channels * 2:
-            return context
-        out_vid = self.video_embeddings_connector(context)[0]
-        out_audio = self.audio_embeddings_connector(context)[0]
-        return torch.concat((out_vid, out_audio), dim=-1)
-
     def _init_transformer_blocks(self, device, dtype, **kwargs):
         """Initialize transformer blocks for LTXAV."""
         self.transformer_blocks = nn.ModuleList(

comfy/ldm/lightricks/embeddings_connector.py

@@ -234,7 +234,7 @@ class Embeddings1DConnector(nn.Module):
 
         return indices
 
-    def precompute_freqs_cis(self, indices_grid, spacing="exp", out_dtype=None):
+    def precompute_freqs_cis(self, indices_grid, spacing="exp"):
         dim = self.inner_dim
         n_elem = 2  # 2 because of cos and sin
         freqs = self.precompute_freqs(indices_grid, spacing)
@@ -247,7 +247,7 @@ class Embeddings1DConnector(nn.Module):
             )
         else:
             cos_freq, sin_freq = interleaved_freqs_cis(freqs, dim % n_elem)
-        return cos_freq.to(dtype=out_dtype), sin_freq.to(dtype=out_dtype), self.split_rope
+        return cos_freq.to(self.dtype), sin_freq.to(self.dtype), self.split_rope
 
     def forward(
         self,
@@ -288,7 +288,7 @@ class Embeddings1DConnector(nn.Module):
             hidden_states.shape[1], dtype=torch.float32, device=hidden_states.device
         )
         indices_grid = indices_grid[None, None, :]
-        freqs_cis = self.precompute_freqs_cis(indices_grid, out_dtype=hidden_states.dtype)
+        freqs_cis = self.precompute_freqs_cis(indices_grid)
 
         # 2. Blocks
         for block_idx, block in enumerate(self.transformer_1d_blocks):

comfy/memory_management.py

@@ -78,4 +78,4 @@ def interpret_gathered_like(tensors, gathered):
 
     return dest_views
 
-aimdo_enabled = False
+aimdo_allocator = None

comfy/model_base.py

@@ -988,14 +988,10 @@ class LTXAV(BaseModel):
     def extra_conds(self, **kwargs):
         out = super().extra_conds(**kwargs)
         attention_mask = kwargs.get("attention_mask", None)
-        device = kwargs["device"]
-
         if attention_mask is not None:
             out['attention_mask'] = comfy.conds.CONDRegular(attention_mask)
         cross_attn = kwargs.get("cross_attn", None)
         if cross_attn is not None:
-            if hasattr(self.diffusion_model, "preprocess_text_embeds"):
-                cross_attn = self.diffusion_model.preprocess_text_embeds(cross_attn.to(device=device, dtype=self.get_dtype_inference()))
             out['c_crossattn'] = comfy.conds.CONDRegular(cross_attn)
 
         out['frame_rate'] = comfy.conds.CONDConstant(kwargs.get("frame_rate", 25))

comfy/model_management.py

@@ -836,7 +836,7 @@ def unet_inital_load_device(parameters, dtype):
 
     mem_dev = get_free_memory(torch_dev)
     mem_cpu = get_free_memory(cpu_dev)
-    if mem_dev > mem_cpu and model_size < mem_dev and comfy.memory_management.aimdo_enabled:
+    if mem_dev > mem_cpu and model_size < mem_dev and comfy.memory_management.aimdo_allocator is None:
         return torch_dev
     else:
         return cpu_dev
@@ -1121,6 +1121,7 @@ def get_cast_buffer(offload_stream, device, size, ref):
             synchronize()
             del STREAM_CAST_BUFFERS[offload_stream]
             del cast_buffer
+            #FIXME: This doesn't work in Aimdo because mempool cant clear cache
             soft_empty_cache()
         with wf_context:
             cast_buffer = torch.empty((size), dtype=torch.int8, device=device)

comfy/text_encoders/lt.py

@@ -3,6 +3,7 @@ import os
 from transformers import T5TokenizerFast
 from .spiece_tokenizer import SPieceTokenizer
 import comfy.text_encoders.genmo
+from comfy.ldm.lightricks.embeddings_connector import Embeddings1DConnector
 import torch
 import comfy.utils
 import math
@@ -108,6 +109,22 @@ class LTXAVTEModel(torch.nn.Module):
         operations = self.gemma3_12b.operations # TODO
         self.text_embedding_projection = operations.Linear(3840 * 49, 3840, bias=False, dtype=dtype, device=device)
 
+        self.audio_embeddings_connector = Embeddings1DConnector(
+            split_rope=True,
+            double_precision_rope=True,
+            dtype=dtype,
+            device=device,
+            operations=operations,
+        )
+
+        self.video_embeddings_connector = Embeddings1DConnector(
+            split_rope=True,
+            double_precision_rope=True,
+            dtype=dtype,
+            device=device,
+            operations=operations,
+        )
+
     def set_clip_options(self, options):
         self.execution_device = options.get("execution_device", self.execution_device)
         self.gemma3_12b.set_clip_options(options)
@@ -129,6 +146,10 @@ class LTXAVTEModel(torch.nn.Module):
         out = out.reshape((out.shape[0], out.shape[1], -1))
         out = self.text_embedding_projection(out)
         out = out.float()
+        out_vid = self.video_embeddings_connector(out)[0]
+        out_audio = self.audio_embeddings_connector(out)[0]
+        out = torch.concat((out_vid, out_audio), dim=-1)
+
         return out.to(out_device), pooled
 
     def generate(self, tokens, do_sample, max_length, temperature, top_k, top_p, min_p, repetition_penalty, seed):
@@ -138,14 +159,14 @@ class LTXAVTEModel(torch.nn.Module):
         if "model.layers.47.self_attn.q_norm.weight" in sd:
             return self.gemma3_12b.load_sd(sd)
         else:
-            sdo = comfy.utils.state_dict_prefix_replace(sd, {"text_embedding_projection.aggregate_embed.weight": "text_embedding_projection.weight"}, filter_keys=True)
+            sdo = comfy.utils.state_dict_prefix_replace(sd, {"text_embedding_projection.aggregate_embed.weight": "text_embedding_projection.weight", "model.diffusion_model.video_embeddings_connector.": "video_embeddings_connector.", "model.diffusion_model.audio_embeddings_connector.": "audio_embeddings_connector."}, filter_keys=True)
             if len(sdo) == 0:
                 sdo = sd
 
             missing_all = []
             unexpected_all = []
 
-            for prefix, component in [("text_embedding_projection.", self.text_embedding_projection)]:
+            for prefix, component in [("text_embedding_projection.", self.text_embedding_projection), ("video_embeddings_connector.", self.video_embeddings_connector), ("audio_embeddings_connector.", self.audio_embeddings_connector)]:
                 component_sd = {k.replace(prefix, ""): v for k, v in sdo.items() if k.startswith(prefix)}
                 if component_sd:
                     missing, unexpected = component.load_state_dict(component_sd, strict=False, assign=getattr(self, "can_assign_sd", False))

comfy/utils.py

@@ -1154,7 +1154,7 @@ def tiled_scale(samples, function, tile_x=64, tile_y=64, overlap = 8, upscale_am
     return tiled_scale_multidim(samples, function, (tile_y, tile_x), overlap=overlap, upscale_amount=upscale_amount, out_channels=out_channels, output_device=output_device, pbar=pbar)
 
 def model_trange(*args, **kwargs):
-    if not comfy.memory_management.aimdo_enabled:
+    if comfy.memory_management.aimdo_allocator is None:
         return trange(*args, **kwargs)
 
     pbar = trange(*args, **kwargs, smoothing=1.0)

comfy_api/latest/_input_impl/video_types.py

@@ -444,7 +444,7 @@ class VideoFromComponents(VideoInput):
             output.mux(packet)
 
             if audio_stream and self.__components.audio:
-                frame = av.AudioFrame.from_ndarray(waveform.float().cpu().contiguous().numpy(), format='fltp', layout=layout)
+                frame = av.AudioFrame.from_ndarray(waveform.float().cpu().numpy(), format='fltp', layout=layout)
                 frame.sample_rate = audio_sample_rate
                 frame.pts = 0
                 output.mux(audio_stream.encode(frame))

comfy_api/latest/_io.py

@@ -73,7 +73,6 @@ class RemoteOptions:
 class NumberDisplay(str, Enum):
     number = "number"
     slider = "slider"
-    gradient_slider = "gradientslider"
 
 
 class ControlAfterGenerate(str, Enum):
@@ -297,15 +296,13 @@ class Float(ComfyTypeIO):
         '''Float input.'''
         def __init__(self, id: str, display_name: str=None, optional=False, tooltip: str=None, lazy: bool=None,
                     default: float=None, min: float=None, max: float=None, step: float=None, round: float=None,
-                    display_mode: NumberDisplay=None, gradient_stops: list[list[float]]=None,
-                    socketless: bool=None, force_input: bool=None, extra_dict=None, raw_link: bool=None, advanced: bool=None):
+                    display_mode: NumberDisplay=None, socketless: bool=None, force_input: bool=None, extra_dict=None, raw_link: bool=None, advanced: bool=None):
             super().__init__(id, display_name, optional, tooltip, lazy, default, socketless, None, force_input, extra_dict, raw_link, advanced)
             self.min = min
             self.max = max
             self.step = step
             self.round = round
             self.display_mode = display_mode
-            self.gradient_stops = gradient_stops
             self.default: float
 
         def as_dict(self):
@@ -315,7 +312,6 @@ class Float(ComfyTypeIO):
                 "step": self.step,
                 "round": self.round,
                 "display": self.display_mode,
-                "gradient_stops": self.gradient_stops,
             })
 
 @comfytype(io_type="STRING")
@@ -1237,82 +1233,6 @@ class BoundingBox(ComfyTypeIO):
             return d
 
 
-@comfytype(io_type="CURVE")
-class Curve(ComfyTypeIO):
-    Type = list
-
-    class Input(WidgetInput):
-        def __init__(self, id: str, display_name: str=None, optional=False, tooltip: str=None,
-                     socketless: bool=True, default: list=None, advanced: bool=None):
-            super().__init__(id, display_name, optional, tooltip, None, default, socketless, None, None, None, None, advanced)
-            if default is None:
-                self.default = [[0, 0], [1, 1]]
-
-        def as_dict(self):
-            return super().as_dict()
-
-
-@comfytype(io_type="RANGE")
-class Range(ComfyTypeIO):
-    Type = dict  # {"min": float, "max": float, "midpoint"?: float}
-
-    class Input(WidgetInput):
-        def __init__(self, id: str, display_name: str=None, optional=False, tooltip: str=None,
-                     socketless: bool=True, default: dict=None,
-                     display: str=None,
-                     gradient_stops: list=None,
-                     show_midpoint: bool=None,
-                     midpoint_scale: str=None,
-                     value_min: float=None,
-                     value_max: float=None,
-                     advanced: bool=None):
-            super().__init__(id, display_name, optional, tooltip, None, default, socketless, None, None, None, None, advanced)
-            if default is None:
-                self.default = {"min": 0.0, "max": 1.0}
-            self.display = display
-            self.gradient_stops = gradient_stops
-            self.show_midpoint = show_midpoint
-            self.midpoint_scale = midpoint_scale
-            self.value_min = value_min
-            self.value_max = value_max
-
-        def as_dict(self):
-            return super().as_dict() | prune_dict({
-                "display": self.display,
-                "gradient_stops": self.gradient_stops,
-                "show_midpoint": self.show_midpoint,
-                "midpoint_scale": self.midpoint_scale,
-                "value_min": self.value_min,
-                "value_max": self.value_max,
-            })
-
-
-@comfytype(io_type="COLOR_CURVES")
-class ColorCurves(ComfyTypeIO):
-    class ColorCurvesDict(TypedDict):
-        rgb: list[list[float]]
-        red: list[list[float]]
-        green: list[list[float]]
-        blue: list[list[float]]
-
-    Type = ColorCurvesDict
-
-    class Input(WidgetInput):
-        def __init__(self, id: str, display_name: str=None, optional=False, tooltip: str=None,
-                     socketless: bool=True, default: dict=None, advanced: bool=None):
-            super().__init__(id, display_name, optional, tooltip, None, default, socketless, None, None, None, None, advanced)
-            if default is None:
-                self.default = {
-                    "rgb": [[0, 0], [1, 1]],
-                    "red": [[0, 0], [1, 1]],
-                    "green": [[0, 0], [1, 1]],
-                    "blue": [[0, 0], [1, 1]]
-                }
-
-        def as_dict(self):
-            return super().as_dict()
-
-
 DYNAMIC_INPUT_LOOKUP: dict[str, Callable[[dict[str, Any], dict[str, Any], tuple[str, dict[str, Any]], str, list[str] | None], None]] = {}
 def register_dynamic_input_func(io_type: str, func: Callable[[dict[str, Any], dict[str, Any], tuple[str, dict[str, Any]], str, list[str] | None], None]):
     DYNAMIC_INPUT_LOOKUP[io_type] = func
@@ -2299,7 +2219,5 @@ __all__ = [
     "PriceBadgeDepends",
     "PriceBadge",
     "BoundingBox",
-    "Curve",
-    "ColorCurves",
     "NodeReplace",
 ]

comfy_extras/nodes_canny.py

@@ -10,7 +10,6 @@ class Canny(io.ComfyNode):
     def define_schema(cls):
         return io.Schema(
             node_id="Canny",
-            display_name="Canny",
             search_aliases=["edge detection", "outline", "contour detection", "line art"],
             category="image/preprocessors",
             essentials_category="Image Tools",

comfy_extras/nodes_color_curves.py

@@ -1,137 +0,0 @@
-from typing_extensions import override
-import torch
-import numpy as np
-
-from comfy_api.latest import ComfyExtension, io, ui
-
-
-def _monotone_cubic_hermite(xs, ys, x_query):
-    """Evaluate monotone cubic Hermite interpolation at x_query points."""
-    n = len(xs)
-    if n == 0:
-        return np.zeros_like(x_query)
-    if n == 1:
-        return np.full_like(x_query, ys[0])
-
-    # Compute slopes
-    deltas = np.diff(ys) / np.maximum(np.diff(xs), 1e-10)
-
-    # Compute tangents (Fritsch-Carlson)
-    slopes = np.zeros(n)
-    slopes[0] = deltas[0]
-    slopes[-1] = deltas[-1]
-    for i in range(1, n - 1):
-        if deltas[i - 1] * deltas[i] <= 0:
-            slopes[i] = 0
-        else:
-            slopes[i] = (deltas[i - 1] + deltas[i]) / 2
-
-    # Enforce monotonicity
-    for i in range(n - 1):
-        if deltas[i] == 0:
-            slopes[i] = 0
-            slopes[i + 1] = 0
-        else:
-            alpha = slopes[i] / deltas[i]
-            beta = slopes[i + 1] / deltas[i]
-            s = alpha ** 2 + beta ** 2
-            if s > 9:
-                t = 3 / np.sqrt(s)
-                slopes[i] = t * alpha * deltas[i]
-                slopes[i + 1] = t * beta * deltas[i]
-
-    # Evaluate
-    result = np.zeros_like(x_query, dtype=np.float64)
-    indices = np.searchsorted(xs, x_query, side='right') - 1
-    indices = np.clip(indices, 0, n - 2)
-
-    for i in range(n - 1):
-        mask = indices == i
-        if not np.any(mask):
-            continue
-        dx = xs[i + 1] - xs[i]
-        if dx == 0:
-            result[mask] = ys[i]
-            continue
-        t = (x_query[mask] - xs[i]) / dx
-        t2 = t * t
-        t3 = t2 * t
-        h00 = 2 * t3 - 3 * t2 + 1
-        h10 = t3 - 2 * t2 + t
-        h01 = -2 * t3 + 3 * t2
-        h11 = t3 - t2
-        result[mask] = h00 * ys[i] + h10 * dx * slopes[i] + h01 * ys[i + 1] + h11 * dx * slopes[i + 1]
-
-    # Clamp edges
-    result[x_query <= xs[0]] = ys[0]
-    result[x_query >= xs[-1]] = ys[-1]
-
-    return result
-
-
-def _build_lut(points):
-    """Build a 256-entry LUT from curve control points in [0,1] space."""
-    if not points or len(points) < 2:
-        return np.arange(256, dtype=np.float64) / 255.0
-
-    pts = sorted(points, key=lambda p: p[0])
-    xs = np.array([p[0] for p in pts], dtype=np.float64)
-    ys = np.array([p[1] for p in pts], dtype=np.float64)
-
-    x_query = np.linspace(0, 1, 256)
-    lut = _monotone_cubic_hermite(xs, ys, x_query)
-    return np.clip(lut, 0, 1)
-
-
-class ColorCurvesNode(io.ComfyNode):
-    @classmethod
-    def define_schema(cls):
-        return io.Schema(
-            node_id="ColorCurves",
-            display_name="Color Curves",
-            category="image/adjustment",
-            inputs=[
-                io.Image.Input("image"),
-                io.ColorCurves.Input("settings"),
-            ],
-            outputs=[
-                io.Image.Output(),
-            ],
-        )
-
-    @classmethod
-    def execute(cls, image: torch.Tensor, settings: dict) -> io.NodeOutput:
-        rgb_pts = settings.get("rgb", [[0, 0], [1, 1]])
-        red_pts = settings.get("red", [[0, 0], [1, 1]])
-        green_pts = settings.get("green", [[0, 0], [1, 1]])
-        blue_pts = settings.get("blue", [[0, 0], [1, 1]])
-
-        rgb_lut = _build_lut(rgb_pts)
-        red_lut = _build_lut(red_pts)
-        green_lut = _build_lut(green_pts)
-        blue_lut = _build_lut(blue_pts)
-
-        # Convert to numpy for LUT application
-        img_np = image.cpu().numpy().copy()
-
-        # Apply per-channel curves then RGB master curve.
-        # Index with floor(val * 256) clamped to [0, 255] to match GPU NEAREST
-        # texture sampling on a 256-wide LUT texture.
-        for ch, ch_lut in enumerate([red_lut, green_lut, blue_lut]):
-            indices = np.clip((img_np[..., ch] * 256).astype(np.int32), 0, 255)
-            img_np[..., ch] = ch_lut[indices]
-            indices = np.clip((img_np[..., ch] * 256).astype(np.int32), 0, 255)
-            img_np[..., ch] = rgb_lut[indices]
-
-        result = torch.from_numpy(np.clip(img_np, 0, 1)).to(image.device, dtype=image.dtype)
-        return io.NodeOutput(result, ui=ui.PreviewImage(result))
-
-
-class ColorCurvesExtension(ComfyExtension):
-    @override
-    async def get_node_list(self) -> list[type[io.ComfyNode]]:
-        return [ColorCurvesNode]
-
-
-async def comfy_entrypoint() -> ColorCurvesExtension:
-    return ColorCurvesExtension()

comfy_extras/nodes_glsl.py

@@ -716,12 +716,12 @@ def _render_shader_batch(
         gl.glBindFramebuffer(gl.GL_FRAMEBUFFER, 0)
         gl.glUseProgram(0)
 
-        for tex in input_textures:
-            gl.glDeleteTextures(tex)
-        for tex in output_textures:
-            gl.glDeleteTextures(tex)
-        for tex in ping_pong_textures:
-            gl.glDeleteTextures(tex)
+        if input_textures:
+            gl.glDeleteTextures(len(input_textures), input_textures)
+        if output_textures:
+            gl.glDeleteTextures(len(output_textures), output_textures)
+        if ping_pong_textures:
+            gl.glDeleteTextures(len(ping_pong_textures), ping_pong_textures)
         if fbo is not None:
             gl.glDeleteFramebuffers(1, [fbo])
         for pp_fbo in ping_pong_fbos:

comfy_extras/nodes_images.py

@@ -588,7 +588,6 @@ class ImageRotate(IO.ComfyNode):
     def define_schema(cls):
         return IO.Schema(
             node_id="ImageRotate",
-            display_name="Image Rotate",
             search_aliases=["turn", "flip orientation"],
             category="image/transform",
             essentials_category="Image Tools",

comfy_extras/nodes_nag.py

@@ -10,7 +10,7 @@ class NAGuidance(io.ComfyNode):
             node_id="NAGuidance",
             display_name="Normalized Attention Guidance",
             description="Applies Normalized Attention Guidance to models, enabling negative prompts on distilled/schnell models.",
-            category="advanced/guidance",
+            category="",
             is_experimental=True,
             inputs=[
                 io.Model.Input("model", tooltip="The model to apply NAG to."),

comfy_extras/nodes_post_processing.py

@@ -19,7 +19,6 @@ class Blend(io.ComfyNode):
     def define_schema(cls):
         return io.Schema(
             node_id="ImageBlend",
-            display_name="Image Blend",
             category="image/postprocessing",
             inputs=[
                 io.Image.Input("image1"),
@@ -77,7 +76,6 @@ class Blur(io.ComfyNode):
     def define_schema(cls):
         return io.Schema(
             node_id="ImageBlur",
-            display_name="Image Blur",
             category="image/postprocessing",
             essentials_category="Image Tools",
             inputs=[

comfy_extras/nodes_primitive.py

@@ -29,7 +29,6 @@ class StringMultiline(io.ComfyNode):
             node_id="PrimitiveStringMultiline",
             display_name="String (Multiline)",
             category="utils/primitive",
-            essentials_category="Basics",
             inputs=[
                 io.String.Input("value", multiline=True),
             ],

cuda_malloc.py

@@ -1,8 +1,10 @@
 import os
 import importlib.util
-from comfy.cli_args import args, PerformanceFeature
+from comfy.cli_args import args, PerformanceFeature, enables_dynamic_vram
 import subprocess
 
+import comfy_aimdo.control
+
 #Can't use pytorch to get the GPU names because the cuda malloc has to be set before the first import.
 def get_gpu_names():
     if os.name == 'nt':
@@ -85,6 +87,10 @@ if not args.cuda_malloc:
     except:
         pass
 
+if enables_dynamic_vram() and comfy_aimdo.control.init():
+    args.cuda_malloc = False
+    os.environ['PYTORCH_CUDA_ALLOC_CONF'] = ""
+
 if args.disable_cuda_malloc:
     args.cuda_malloc = False
 

execution.py

@@ -9,6 +9,7 @@ import traceback
 from enum import Enum
 from typing import List, Literal, NamedTuple, Optional, Union
 import asyncio
+from contextlib import nullcontext
 
 import torch
 
@@ -520,14 +521,19 @@ async def execute(server, dynprompt, caches, current_item, extra_data, executed,
                 # TODO - How to handle this with async functions without contextvars (which requires Python 3.12)?
                 GraphBuilder.set_default_prefix(unique_id, call_index, 0)
 
-            try:
-                output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
-            finally:
-                if comfy.memory_management.aimdo_enabled:
-                    if args.verbose == "DEBUG":
-                        comfy_aimdo.control.analyze()
-                    comfy.model_management.reset_cast_buffers()
-                    comfy_aimdo.model_vbar.vbars_reset_watermark_limits()
+            #Do comfy_aimdo mempool chunking here on the per-node level. Multi-model workflows
+            #will cause all sorts of incompatible memory shapes to fragment the pytorch alloc
+            #that we just want to cull out each model run.
+            allocator = comfy.memory_management.aimdo_allocator
+            with nullcontext() if allocator is None else torch.cuda.use_mem_pool(torch.cuda.MemPool(allocator.allocator())):
+                try:
+                    output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
+                finally:
+                    if allocator is not None:
+                        if args.verbose == "DEBUG":
+                            comfy_aimdo.model_vbar.vbars_analyze()
+                        comfy.model_management.reset_cast_buffers()
+                        comfy_aimdo.model_vbar.vbars_reset_watermark_limits()
 
             if has_pending_tasks:
                 pending_async_nodes[unique_id] = output_data

main.py

@@ -173,10 +173,6 @@ import gc
 if 'torch' in sys.modules:
     logging.warning("WARNING: Potential Error in code: Torch already imported, torch should never be imported before this point.")
 
-import comfy_aimdo.control
-
-if enables_dynamic_vram():
-    comfy_aimdo.control.init()
 
 import comfy.utils
 
@@ -192,9 +188,13 @@ import hook_breaker_ac10a0
 import comfy.memory_management
 import comfy.model_patcher
 
+import comfy_aimdo.control
+import comfy_aimdo.torch
+
 if enables_dynamic_vram():
     if comfy.model_management.torch_version_numeric < (2, 8):
         logging.warning("Unsupported Pytorch detected. DynamicVRAM support requires Pytorch version 2.8 or later. Falling back to legacy ModelPatcher. VRAM estimates may be unreliable especially on Windows")
+        comfy.memory_management.aimdo_allocator = None
     elif comfy_aimdo.control.init_device(comfy.model_management.get_torch_device().index):
         if args.verbose == 'DEBUG':
             comfy_aimdo.control.set_log_debug()
@@ -208,10 +208,11 @@ if enables_dynamic_vram():
             comfy_aimdo.control.set_log_info()
 
         comfy.model_patcher.CoreModelPatcher = comfy.model_patcher.ModelPatcherDynamic
-        comfy.memory_management.aimdo_enabled = True
+        comfy.memory_management.aimdo_allocator = comfy_aimdo.torch.get_torch_allocator()
         logging.info("DynamicVRAM support detected and enabled")
     else:
         logging.warning("No working comfy-aimdo install detected. DynamicVRAM support disabled. Falling back to legacy ModelPatcher. VRAM estimates may be unreliable especially on Windows")
+        comfy.memory_management.aimdo_allocator = None
 
 
 def cuda_malloc_warning():

nodes.py

@@ -70,6 +70,7 @@ class CLIPTextEncode(ComfyNodeABC):
     FUNCTION = "encode"
 
     CATEGORY = "conditioning"
+    ESSENTIALS_CATEGORY = "Basics"
     DESCRIPTION = "Encodes a text prompt using a CLIP model into an embedding that can be used to guide the diffusion model towards generating specific images."
     SEARCH_ALIASES = ["text", "prompt", "text prompt", "positive prompt", "negative prompt", "encode text", "text encoder", "encode prompt"]
 
@@ -2035,144 +2036,6 @@ class ImagePadForOutpaint:
         return (new_image, mask.unsqueeze(0))
 
 
-class TestCurveWidget:
-    @classmethod
-    def INPUT_TYPES(s):
-        return {
-            "required": {
-                "curve": ("CURVE", {"default": [[0, 0], [1, 1]]}),
-            }
-        }
-
-    RETURN_TYPES = ("STRING",)
-    RETURN_NAMES = ("points",)
-    FUNCTION = "execute"
-    OUTPUT_NODE = True
-    CATEGORY = "testing"
-
-    def execute(self, curve):
-        import json
-        result = json.dumps(curve, indent=2)
-        print("Curve points:", result)
-        return {"ui": {"text": [result]}, "result": (result,)}
-
-
-class TestRangePlain:
-    @classmethod
-    def INPUT_TYPES(s):
-        return {
-            "required": {
-                "range": ("RANGE", {"default": {"min": 0.0, "max": 1.0}}),
-                "range_midpoint": ("RANGE", {
-                    "default": {"min": 0.2, "max": 0.8, "midpoint": 0.5},
-                    "show_midpoint": True,
-                }),
-            }
-        }
-
-    RETURN_TYPES = ("STRING",)
-    FUNCTION = "execute"
-    OUTPUT_NODE = True
-    CATEGORY = "testing"
-
-    def execute(self, **kwargs):
-        import json
-        result = json.dumps(kwargs, indent=2)
-        return {"ui": {"text": [result]}, "result": (result,)}
-
-
-class TestRangeGradient:
-    @classmethod
-    def INPUT_TYPES(s):
-        return {
-            "required": {
-                "range": ("RANGE", {
-                    "default": {"min": 0.0, "max": 1.0},
-                    "display": "gradient",
-                    "gradient_stops": [
-                        {"offset": 0.0, "color": [0, 0, 0]},
-                        {"offset": 1.0, "color": [255, 255, 255]}
-                    ],
-                }),
-                "range_midpoint": ("RANGE", {
-                    "default": {"min": 0.0, "max": 1.0, "midpoint": 0.5},
-                    "display": "gradient",
-                    "gradient_stops": [
-                        {"offset": 0.0, "color": [0, 0, 0]},
-                        {"offset": 1.0, "color": [255, 255, 255]}
-                    ],
-                    "show_midpoint": True,
-                    "midpoint_scale": "gamma",
-                }),
-            }
-        }
-
-    RETURN_TYPES = ("STRING",)
-    FUNCTION = "execute"
-    OUTPUT_NODE = True
-    CATEGORY = "testing"
-
-    def execute(self, **kwargs):
-        import json
-        result = json.dumps(kwargs, indent=2)
-        return {"ui": {"text": [result]}, "result": (result,)}
-
-
-class TestRangeHistogram:
-    RANGE_OPTS = {
-        "display": "histogram",
-        "show_midpoint": True,
-        "midpoint_scale": "gamma",
-        "value_min": 0,
-        "value_max": 255,
-    }
-
-    @classmethod
-    def INPUT_TYPES(s):
-        default = {"min": 0, "max": 255, "midpoint": 0.5}
-        return {
-            "required": {
-                "image": ("IMAGE",),
-                "rgb":   ("RANGE", {"default": {**default}, **s.RANGE_OPTS}),
-                "red":   ("RANGE", {"default": {**default}, **s.RANGE_OPTS}),
-                "green": ("RANGE", {"default": {**default}, **s.RANGE_OPTS}),
-                "blue":  ("RANGE", {"default": {**default}, **s.RANGE_OPTS}),
-            }
-        }
-
-    RETURN_TYPES = ("STRING",)
-    FUNCTION = "execute"
-    OUTPUT_NODE = True
-    CATEGORY = "testing"
-
-    def execute(self, image, rgb, red, green, blue):
-        import json
-        import numpy as np
-
-        img = image[0].cpu().numpy()  # (H, W, C)
-
-        # Per-channel histograms
-        hist_r, _ = np.histogram(img[:, :, 0].flatten(), bins=256, range=(0.0, 1.0))
-        hist_g, _ = np.histogram(img[:, :, 1].flatten(), bins=256, range=(0.0, 1.0))
-        hist_b, _ = np.histogram(img[:, :, 2].flatten(), bins=256, range=(0.0, 1.0))
-
-        # Luminance histogram (BT.709)
-        luminance = 0.2126 * img[:, :, 0] + 0.7152 * img[:, :, 1] + 0.0722 * img[:, :, 2]
-        hist_rgb, _ = np.histogram(luminance.flatten(), bins=256, range=(0.0, 1.0))
-
-        result = json.dumps({"rgb": rgb, "red": red, "green": green, "blue": blue}, indent=2)
-        return {
-            "ui": {
-                "text": [result],
-                "range_histogram_rgb":   hist_rgb.astype(np.uint32).tolist(),
-                "range_histogram_red":   hist_r.astype(np.uint32).tolist(),
-                "range_histogram_green": hist_g.astype(np.uint32).tolist(),
-                "range_histogram_blue":  hist_b.astype(np.uint32).tolist(),
-            },
-            "result": (result,)
-        }
-
-
 NODE_CLASS_MAPPINGS = {
     "KSampler": KSampler,
     "CheckpointLoaderSimple": CheckpointLoaderSimple,
@@ -2241,10 +2104,6 @@ NODE_CLASS_MAPPINGS = {
     "ConditioningZeroOut": ConditioningZeroOut,
     "ConditioningSetTimestepRange": ConditioningSetTimestepRange,
     "LoraLoaderModelOnly": LoraLoaderModelOnly,
-    "TestCurveWidget": TestCurveWidget,
-    "TestRangePlain": TestRangePlain,
-    "TestRangeGradient": TestRangeGradient,
-    "TestRangeHistogram": TestRangeHistogram,
 }
 
 NODE_DISPLAY_NAME_MAPPINGS = {
@@ -2313,10 +2172,6 @@ NODE_DISPLAY_NAME_MAPPINGS = {
     # _for_testing
     "VAEDecodeTiled": "VAE Decode (Tiled)",
     "VAEEncodeTiled": "VAE Encode (Tiled)",
-    "TestCurveWidget": "Test Curve Widget",
-    "TestRangePlain": "Test Range (Plain)",
-    "TestRangeGradient": "Test Range (Gradient)",
-    "TestRangeHistogram": "Test Range (Histogram)",
 }
 
 EXTENSION_WEB_DIRS = {}

requirements.txt

@@ -22,7 +22,7 @@ alembic
 SQLAlchemy
 av>=14.2.0
 comfy-kitchen>=0.2.7
-comfy-aimdo>=0.2.0
+comfy-aimdo>=0.1.8
 requests
 
 #non essential dependencies: