From 9920249d0fc6e57cc666b6b0efd5e645541094cd Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 06:36:57 +0530
Subject: [PATCH 01/20] Update README.md
---
README.md | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/README.md b/README.md
index 35af370..74bcd3c 100644
--- a/README.md
+++ b/README.md
@@ -24,7 +24,8 @@ The Flux Resolution Calculator is designed to determine the optimal image resolu
- **Supported Megapixels:** 0.1 MP, 1.0 MP, 2.0 MP, 2.1 MP, 2.2 MP, 2.3 MP, 2.4MP, 2.5MP
- **Note:** Generations above 1 MP may appear slightly blurry, but resolutions of 3k+ have been successfully tested on the Flux1.Dev model.
-- **Custom Ratio:** Custom Ratio is now supported. Enable or Disable Custom Ratio and input any ratio. (Example: 4:9)
+- **Custom Ratio:** Custom Ratio is now supported. Enable or Disable Custom Ratio and input any ratio. (Example: 4:9).
+- **Preview:** The preview node is just a visual representation of the ratio.
### Flux Sampler
@@ -35,7 +36,9 @@ The Flux Sampler node combines the functionality of the CustomSamplerAdvance nod
- **Compatibility:** Only the samplers and schedulers compatible with the Flux model are included.
- **Latent Compatibility:** Use SD3 Empty Latent Image only. The normal empty latent image node is not compatible.
-
+
+
+
### Flux Union ControlNet Apply
From ad2c39a187cc4194d3e923bfd9e095827562690e Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 10:09:13 +0530
Subject: [PATCH 02/20] Update README.md
---
README.md | 88 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 88 insertions(+)
diff --git a/README.md b/README.md
index 74bcd3c..89a7c08 100644
--- a/README.md
+++ b/README.md
@@ -17,6 +17,14 @@ This repository contains custom nodes designed for the ComfyUI framework, focusi
- Image
- Get Image Size & Ratio
- Noise Plus Blend
+- Flux Region
+ - Region Mask Generator
+ - Region Mask Processor
+ - Region Mask Validator
+ - Region Mask Conditioning
+ - Flux Attention Control
+ - Region Overlay Visualizer
+ - Flux Attention Cleanup
### Flux Resolution Calculator
@@ -176,6 +184,86 @@ Without Noise Blend:
With Noise Blend:
+### Flux Region (Spatial Control)
+
+The node pipeline is as follows: Region Mask Generator --> Region Mask Processor --> Region Mask Validator --> Flux Region Conditioning --> Flux Attention Control --> Flux Overlay Visualizer (optional) --> Flux Attention Cleanup.
+*Note: Watching the video tutorial is a must. The learning curve is a bit high to use Flux Region Spatial Control*
+
+**Region Mask Generator:** This node generates the regions in mask and bbox format. This information is then passed on to the Mask Processor.
+
+
+
+**Region Mask Processor:** This node generates, processes the mask, and applies Gaussian Blue and feathering. This pre-processor node preprocesses the mask and sends the preprocessed information in the pipeline.
+
+
+
+**Region Mask Validator:** This node calculates the validity of the regions. The "is valid" message will be true if there are no overlaps. The validation message would show you detailed information on the overlapping regions and the overlap percentage. Although the methodology used requires zero overlaps, the issue is resolved in the flux attention control with feathering. Overlapping will only be an issue if it is excessive.
+
+
+
+**Region Mask Conditioning:** Up to three separate conditioning can be connected. The node will process based on the number of regions defined rather than the actual conditioning connections. The strength values are independent for each region. Strength 1 for Region 1, Strength 2 for Region 2, and Strength 3 for Region 3. The strength value range is from 0 to 10 with an increment/decrement step of 0.1. At Value 1, the region strength will match the base conditioning strength, which is always set at 1 as a global value. Strength Values are not only relative to the base conditioning values but are also relative to each other. They are also affected by the Region % area in the canvas and the feathering value in the attention control. Please note. Only use the dual clip and flux conditioning in comfy. The base + region flux guidance should be set to 1.
+
+
+
+**Flux Attention Control:** The node takes the region conditioning + base conditioning + the feathering strengths and all the previous information in the pipeline and overrides the Flux Attention. When disabled, it only passes through the base conditioning to the Sampler.
+
+
+
+**Region Overlay Visualizer:** This node overlays the region on the final output for visual purposes only.
+
+**Flux Attention Cleanup:** Since the attention is overridden in the model, a tensor mismatch error will occur when you switch the workflow. We also do not want the attention to be cleaned up in the existing workflow. This node automatically will preserve attention during re-runs in the existing workflow, but when switching workflow will do a fresh clean up and restore flux original attention. This process is achieved without a model unload or manual cache cleanup, as they will not work.
+
+
+
+**Xformers & Token Limits:** The pipeline here needs VRAM to be stored in the same GPU while Flux, LoRA, and Dual Clip are loaded. Due to this, XFormers is required to optimize VRAM Usage. The total tensor size is approximately 5000x5000 for a 24GB VRAM GPU. This limit is set on purpose for consumer-grade hardware and will be increased with the release of a 5090 GPU. Due to this tensor size limitation, token size is also limited when using Dual Clip. Single Clip is not recommended. Use only Dual Clip. When Base Plus all 3 Regions are connected, a max token for each clip l and t5xxl should be 80 or less. 160 Tokens when combined. In other words, you can use (160 x 4 = 640) tokens in total before running into a tensor size error. Reduce prompt tokens when you get a tensor size error. This system was implemented to avoid any OOM errors (out-of-memory errors) on a 24GB VRAM GPU. All testing has been done on Flux 1.dev model with FP8 and t5xxl with FP 16 without any OOM on a 4090.
+
+**GGUF & CivitAI fine-tune models:** The Flux Region Pipeline was tested with GGUF models without issues. Third-party CivitAI Copax Timeless XPlus 3 Flux models also worked without problems.
+
+**LoRA Support:** LoRA is supported and will apply to all attention. At this stage, using different LoRA for different Regions is not possible. Research work is still ongoing.
+
+Results:
+**Example 1**
+3 Region Split Blend using Advance LLM: Base Conditioning (ignored) + 3 Regions
+
+
+
+**Example 2**
+Style manipulation: Base Conditioning + 1 Region
+
+
+
+**Example 3**
+Simple Splitting Contrast: Base Conditioning (ignored) + 2 Regions
+
+
+
+**Example 4**
+Simple Splitting Blend: Base Conditioning + 1 Region
+
+
+
+**Example 5**
+3 Region Split Blend: Base Conditioning (ignored) + 3 Regions
+
+
+
+**Example 6**
+3 Region Split Contrast: Base Conditioning (ignored) + 3 Regions
+
+
+
+**Example 7**
+3 Region Split Blend using Advance LLM: Base Conditioning (ignored) + 3 Regions
+
+
+
+**Example 8**
+Color Manipulation: Base Conditioning (ignored) + 3 Regions
+
+
+
+**YouTube tutorial Flux Region Usage: Coming Soon**
+
### YouTube ComfyUI Tutorials
We are a team of two and create extensive tutorials for ComfyUI. Check out our YouTube channel:
From 5ab6d7ffb35c822631edbd243399e15973e089b1 Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 10:11:40 +0530
Subject: [PATCH 03/20] Update README.md
---
README.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/README.md b/README.md
index 89a7c08..e8a476c 100644
--- a/README.md
+++ b/README.md
@@ -187,7 +187,7 @@ With Noise Blend:
### Flux Region (Spatial Control)
The node pipeline is as follows: Region Mask Generator --> Region Mask Processor --> Region Mask Validator --> Flux Region Conditioning --> Flux Attention Control --> Flux Overlay Visualizer (optional) --> Flux Attention Cleanup.
-*Note: Watching the video tutorial is a must. The learning curve is a bit high to use Flux Region Spatial Control*
+*Note: Watching the video tutorial is a must. The learning curve is a bit high to use Flux Region Spatial Control.*
**Region Mask Generator:** This node generates the regions in mask and bbox format. This information is then passed on to the Mask Processor.
From 4fa620b2b03747b7d6b8047e32ff1e0ffad2c12e Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 10:14:02 +0530
Subject: [PATCH 04/20] Update README.md
---
README.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/README.md b/README.md
index e8a476c..a697990 100644
--- a/README.md
+++ b/README.md
@@ -213,7 +213,7 @@ The node pipeline is as follows: Region Mask Generator --> Region Mask Processor
**Flux Attention Cleanup:** Since the attention is overridden in the model, a tensor mismatch error will occur when you switch the workflow. We also do not want the attention to be cleaned up in the existing workflow. This node automatically will preserve attention during re-runs in the existing workflow, but when switching workflow will do a fresh clean up and restore flux original attention. This process is achieved without a model unload or manual cache cleanup, as they will not work.
-
+
**Xformers & Token Limits:** The pipeline here needs VRAM to be stored in the same GPU while Flux, LoRA, and Dual Clip are loaded. Due to this, XFormers is required to optimize VRAM Usage. The total tensor size is approximately 5000x5000 for a 24GB VRAM GPU. This limit is set on purpose for consumer-grade hardware and will be increased with the release of a 5090 GPU. Due to this tensor size limitation, token size is also limited when using Dual Clip. Single Clip is not recommended. Use only Dual Clip. When Base Plus all 3 Regions are connected, a max token for each clip l and t5xxl should be 80 or less. 160 Tokens when combined. In other words, you can use (160 x 4 = 640) tokens in total before running into a tensor size error. Reduce prompt tokens when you get a tensor size error. This system was implemented to avoid any OOM errors (out-of-memory errors) on a 24GB VRAM GPU. All testing has been done on Flux 1.dev model with FP8 and t5xxl with FP 16 without any OOM on a 4090.
From f404fa7f7c5a5da9c63557382114f9e89cfe9784 Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 10:16:17 +0530
Subject: [PATCH 05/20] Update README.md
---
README.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/README.md b/README.md
index a697990..a21f3d9 100644
--- a/README.md
+++ b/README.md
@@ -193,7 +193,7 @@ The node pipeline is as follows: Region Mask Generator --> Region Mask Processor
-**Region Mask Processor:** This node generates, processes the mask, and applies Gaussian Blue and feathering. This pre-processor node preprocesses the mask and sends the preprocessed information in the pipeline.
+**Region Mask Processor:** This node processes the generated mask and applies Gaussian Blue and feathering. This pre-processor node preprocesses the mask and sends the preprocessed information in the pipeline.
From 881ff707ed4b8da65cff36268e6ab7fae2f5ab9c Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 10:17:03 +0530
Subject: [PATCH 06/20] Update README.md
---
README.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/README.md b/README.md
index a21f3d9..c009910 100644
--- a/README.md
+++ b/README.md
@@ -193,7 +193,7 @@ The node pipeline is as follows: Region Mask Generator --> Region Mask Processor
-**Region Mask Processor:** This node processes the generated mask and applies Gaussian Blue and feathering. This pre-processor node preprocesses the mask and sends the preprocessed information in the pipeline.
+**Region Mask Processor:** This node processes the generated mask and applies Gaussian Blur and feathering. This pre-processor node preprocesses the mask and sends the preprocessed information in the pipeline.
From 5babe06a4b3b3f0f431a76cd240940952dfc873a Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 10:18:54 +0530
Subject: [PATCH 07/20] Update README.md
---
README.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/README.md b/README.md
index c009910..2e86f4b 100644
--- a/README.md
+++ b/README.md
@@ -197,7 +197,7 @@ The node pipeline is as follows: Region Mask Generator --> Region Mask Processor
-**Region Mask Validator:** This node calculates the validity of the regions. The "is valid" message will be true if there are no overlaps. The validation message would show you detailed information on the overlapping regions and the overlap percentage. Although the methodology used requires zero overlaps, the issue is resolved in the flux attention control with feathering. Overlapping will only be an issue if it is excessive.
+**Region Mask Validator:** This node calculates the validity of the regions. The "is valid" message will be true if there are no overlaps. The validation message would show you detailed information on the overlapping regions and the overlap percentage. Although the methodology used requires zero overlaps, the issue is resolved in the flux attention control with feathering. Overlapping will only be an issue if it is excessive, beyond 40-50%.
From 192e57452b1cdc2b8f8ed1af3cc6438ffd553ea2 Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 10:20:52 +0530
Subject: [PATCH 08/20] Update README.md
---
README.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/README.md b/README.md
index 2e86f4b..368a999 100644
--- a/README.md
+++ b/README.md
@@ -201,7 +201,7 @@ The node pipeline is as follows: Region Mask Generator --> Region Mask Processor
-**Region Mask Conditioning:** Up to three separate conditioning can be connected. The node will process based on the number of regions defined rather than the actual conditioning connections. The strength values are independent for each region. Strength 1 for Region 1, Strength 2 for Region 2, and Strength 3 for Region 3. The strength value range is from 0 to 10 with an increment/decrement step of 0.1. At Value 1, the region strength will match the base conditioning strength, which is always set at 1 as a global value. Strength Values are not only relative to the base conditioning values but are also relative to each other. They are also affected by the Region % area in the canvas and the feathering value in the attention control. Please note. Only use the dual clip and flux conditioning in comfy. The base + region flux guidance should be set to 1.
+**Region Mask Conditioning:** Up to three separate conditioning can be connected. The node will process based on the number of regions defined rather than the actual conditioning connections. The strength values are independent for each region. Strength 1 for Region 1, Strength 2 for Region 2, and Strength 3 for Region 3. The strength value range is from 0 to 10 with an increment/decrement step of 0.1. At Value 1, the region strength will match the base conditioning strength, which is always set at 1 as a global value. Strength Values are not only relative to the base conditioning value but are also relative to each other. They are also affected by the Region % area in the canvas and the feathering value in the attention control. Please note. Only use the dual clip and flux conditioning in comfy. The base + region flux guidance should be set to 1.
From 28b3bca5599c493a6d72e5255b6f2831b82fca17 Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 10:23:40 +0530
Subject: [PATCH 09/20] Update README.md
---
README.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/README.md b/README.md
index 368a999..6a42b5d 100644
--- a/README.md
+++ b/README.md
@@ -213,7 +213,7 @@ The node pipeline is as follows: Region Mask Generator --> Region Mask Processor
**Flux Attention Cleanup:** Since the attention is overridden in the model, a tensor mismatch error will occur when you switch the workflow. We also do not want the attention to be cleaned up in the existing workflow. This node automatically will preserve attention during re-runs in the existing workflow, but when switching workflow will do a fresh clean up and restore flux original attention. This process is achieved without a model unload or manual cache cleanup, as they will not work.
-
+
**Xformers & Token Limits:** The pipeline here needs VRAM to be stored in the same GPU while Flux, LoRA, and Dual Clip are loaded. Due to this, XFormers is required to optimize VRAM Usage. The total tensor size is approximately 5000x5000 for a 24GB VRAM GPU. This limit is set on purpose for consumer-grade hardware and will be increased with the release of a 5090 GPU. Due to this tensor size limitation, token size is also limited when using Dual Clip. Single Clip is not recommended. Use only Dual Clip. When Base Plus all 3 Regions are connected, a max token for each clip l and t5xxl should be 80 or less. 160 Tokens when combined. In other words, you can use (160 x 4 = 640) tokens in total before running into a tensor size error. Reduce prompt tokens when you get a tensor size error. This system was implemented to avoid any OOM errors (out-of-memory errors) on a 24GB VRAM GPU. All testing has been done on Flux 1.dev model with FP8 and t5xxl with FP 16 without any OOM on a 4090.
From 437fbcace1610eef1aaa15584f041d8d5105a42e Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 10:24:06 +0530
Subject: [PATCH 10/20] Update README.md
---
README.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/README.md b/README.md
index 6a42b5d..368a999 100644
--- a/README.md
+++ b/README.md
@@ -213,7 +213,7 @@ The node pipeline is as follows: Region Mask Generator --> Region Mask Processor
**Flux Attention Cleanup:** Since the attention is overridden in the model, a tensor mismatch error will occur when you switch the workflow. We also do not want the attention to be cleaned up in the existing workflow. This node automatically will preserve attention during re-runs in the existing workflow, but when switching workflow will do a fresh clean up and restore flux original attention. This process is achieved without a model unload or manual cache cleanup, as they will not work.
-
+
**Xformers & Token Limits:** The pipeline here needs VRAM to be stored in the same GPU while Flux, LoRA, and Dual Clip are loaded. Due to this, XFormers is required to optimize VRAM Usage. The total tensor size is approximately 5000x5000 for a 24GB VRAM GPU. This limit is set on purpose for consumer-grade hardware and will be increased with the release of a 5090 GPU. Due to this tensor size limitation, token size is also limited when using Dual Clip. Single Clip is not recommended. Use only Dual Clip. When Base Plus all 3 Regions are connected, a max token for each clip l and t5xxl should be 80 or less. 160 Tokens when combined. In other words, you can use (160 x 4 = 640) tokens in total before running into a tensor size error. Reduce prompt tokens when you get a tensor size error. This system was implemented to avoid any OOM errors (out-of-memory errors) on a 24GB VRAM GPU. All testing has been done on Flux 1.dev model with FP8 and t5xxl with FP 16 without any OOM on a 4090.
From 22982763073cff6636baa2f31bca1076cc0612d1 Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 10:24:40 +0530
Subject: [PATCH 11/20] Update README.md
---
README.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/README.md b/README.md
index 368a999..d8081b7 100644
--- a/README.md
+++ b/README.md
@@ -213,7 +213,7 @@ The node pipeline is as follows: Region Mask Generator --> Region Mask Processor
**Flux Attention Cleanup:** Since the attention is overridden in the model, a tensor mismatch error will occur when you switch the workflow. We also do not want the attention to be cleaned up in the existing workflow. This node automatically will preserve attention during re-runs in the existing workflow, but when switching workflow will do a fresh clean up and restore flux original attention. This process is achieved without a model unload or manual cache cleanup, as they will not work.
-
+
**Xformers & Token Limits:** The pipeline here needs VRAM to be stored in the same GPU while Flux, LoRA, and Dual Clip are loaded. Due to this, XFormers is required to optimize VRAM Usage. The total tensor size is approximately 5000x5000 for a 24GB VRAM GPU. This limit is set on purpose for consumer-grade hardware and will be increased with the release of a 5090 GPU. Due to this tensor size limitation, token size is also limited when using Dual Clip. Single Clip is not recommended. Use only Dual Clip. When Base Plus all 3 Regions are connected, a max token for each clip l and t5xxl should be 80 or less. 160 Tokens when combined. In other words, you can use (160 x 4 = 640) tokens in total before running into a tensor size error. Reduce prompt tokens when you get a tensor size error. This system was implemented to avoid any OOM errors (out-of-memory errors) on a 24GB VRAM GPU. All testing has been done on Flux 1.dev model with FP8 and t5xxl with FP 16 without any OOM on a 4090.
From c9e74fa7cb7560a5fd8f8faf48f1c682f29f0010 Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 10:27:32 +0530
Subject: [PATCH 12/20] Update README.md
---
README.md | 2 ++
1 file changed, 2 insertions(+)
diff --git a/README.md b/README.md
index d8081b7..f15a783 100644
--- a/README.md
+++ b/README.md
@@ -221,6 +221,8 @@ The node pipeline is as follows: Region Mask Generator --> Region Mask Processor
**LoRA Support:** LoRA is supported and will apply to all attention. At this stage, using different LoRA for different Regions is not possible. Research work is still ongoing.
+**ControlNet Support:** Currently not tested. Research work is still ongoing.
+
Results:
**Example 1**
3 Region Split Blend using Advance LLM: Base Conditioning (ignored) + 3 Regions
From 58a2863627b3acc36d2311edec37ed823ed31983 Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 10:30:07 +0530
Subject: [PATCH 13/20] Update README.md
---
README.md | 7 +------
1 file changed, 1 insertion(+), 6 deletions(-)
diff --git a/README.md b/README.md
index f15a783..c8ef136 100644
--- a/README.md
+++ b/README.md
@@ -250,16 +250,11 @@ Simple Splitting Blend: Base Conditioning + 1 Region
**Example 6**
-3 Region Split Contrast: Base Conditioning (ignored) + 3 Regions
-
-
-
-**Example 7**
3 Region Split Blend using Advance LLM: Base Conditioning (ignored) + 3 Regions
-**Example 8**
+**Example 7**
Color Manipulation: Base Conditioning (ignored) + 3 Regions
From 9d02f337a1a78b3b237c21d45fe9d5259c08a3f4 Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 10:34:05 +0530
Subject: [PATCH 14/20] Update README.md
---
README.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/README.md b/README.md
index c8ef136..d7854f6 100644
--- a/README.md
+++ b/README.md
@@ -255,7 +255,7 @@ Simple Splitting Blend: Base Conditioning + 1 Region
**Example 7**
-Color Manipulation: Base Conditioning (ignored) + 3 Regions
+Color Manipulation: Base Conditioning (ignored) + 2 Regions
From 0c90b8420cdf56b21e5924927d2b982a6dac9464 Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 10:37:00 +0530
Subject: [PATCH 15/20] Update README.md
---
README.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/README.md b/README.md
index d7854f6..fd6363b 100644
--- a/README.md
+++ b/README.md
@@ -87,7 +87,7 @@ Results: (Canny and Depth Examples not included. They are straightforward.)
<
-
+
From cb9098edc7401d4c74f2ae1845ec70577978dee2 Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 13:01:40 +0530
Subject: [PATCH 16/20] Update README.md
---
README.md | 2 ++
1 file changed, 2 insertions(+)
diff --git a/README.md b/README.md
index fd6363b..3fd727c 100644
--- a/README.md
+++ b/README.md
@@ -1,3 +1,5 @@
+#Requirements Update: 8 Dec 2024: XFormers is now needed. You can install Xformers without breaking dependencies. PLease follow instructions here: https://github.com/gseth/ControlAltAI-Nodes/issues/9#issuecomment-2525463302
+
# ComfyUI ControlAltAI Nodes
This repository contains custom nodes designed for the ComfyUI framework, focusing on quality-of-life improvements. These nodes aim to make tasks easier and more efficient. Two Flux nodes are available to enhance functionality and streamline workflows within ComfyUI.
From ed2abf89e144fccba75c259514ea76bec8e84bcc Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 13:01:59 +0530
Subject: [PATCH 17/20] Update README.md
---
README.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/README.md b/README.md
index 3fd727c..a508b3e 100644
--- a/README.md
+++ b/README.md
@@ -1,4 +1,4 @@
-#Requirements Update: 8 Dec 2024: XFormers is now needed. You can install Xformers without breaking dependencies. PLease follow instructions here: https://github.com/gseth/ControlAltAI-Nodes/issues/9#issuecomment-2525463302
+#Requirements Update: 8 Dec 2024: XFormers is now needed. You can install Xformers without breaking dependencies. Please follow instructions here: https://github.com/gseth/ControlAltAI-Nodes/issues/9#issuecomment-2525463302
# ComfyUI ControlAltAI Nodes
From 583f83ea4288f3d3f6984afb789d529497debe89 Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 13:18:58 +0530
Subject: [PATCH 18/20] Update README.md
---
README.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/README.md b/README.md
index a508b3e..030e02b 100644
--- a/README.md
+++ b/README.md
@@ -1,4 +1,4 @@
-#Requirements Update: 8 Dec 2024: XFormers is now needed. You can install Xformers without breaking dependencies. Please follow instructions here: https://github.com/gseth/ControlAltAI-Nodes/issues/9#issuecomment-2525463302
+#Requirements Update: 8 Dec 2024: XFormers is now needed. You can install Xformers without breaking dependencies. Please follow the instructions here: https://github.com/gseth/ControlAltAI-Nodes/issues/9#issuecomment-2525463302
# ComfyUI ControlAltAI Nodes
From be3b3d726450d573b1769a4359adb117189eadbc Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 13:19:43 +0530
Subject: [PATCH 19/20] Update README.md
---
README.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/README.md b/README.md
index 030e02b..fc381f6 100644
--- a/README.md
+++ b/README.md
@@ -1,4 +1,4 @@
-#Requirements Update: 8 Dec 2024: XFormers is now needed. You can install Xformers without breaking dependencies. Please follow the instructions here: https://github.com/gseth/ControlAltAI-Nodes/issues/9#issuecomment-2525463302
+## Requirements Update: 8 Dec 2024: XFormers is now needed. You can install Xformers without breaking dependencies. Please follow the instructions here: https://github.com/gseth/ControlAltAI-Nodes/issues/9#issuecomment-2525463302
# ComfyUI ControlAltAI Nodes
From 33f9979884c617e794594855e1c114fd4a1342f7 Mon Sep 17 00:00:00 2001
From: ControlAltAI <6172927+gseth@users.noreply.github.com>
Date: Sun, 8 Dec 2024 13:20:22 +0530
Subject: [PATCH 20/20] Update README.md
---
README.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/README.md b/README.md
index fc381f6..772d274 100644
--- a/README.md
+++ b/README.md
@@ -1,4 +1,4 @@
-## Requirements Update: 8 Dec 2024: XFormers is now needed. You can install Xformers without breaking dependencies. Please follow the instructions here: https://github.com/gseth/ControlAltAI-Nodes/issues/9#issuecomment-2525463302
+### Requirements Update: 8 Dec 2024: XFormers is now needed. You can install Xformers without breaking dependencies. Please follow the instructions here: https://github.com/gseth/ControlAltAI-Nodes/issues/9#issuecomment-2525463302
# ComfyUI ControlAltAI Nodes