Workflows and documentation

workflows/README.md

@@ -0,0 +1,39 @@
+# Workflows
+
+
+## Simple sdxl turbo workflow
+
+![turbo workflow](https://raw.githubusercontent.com/Lerc/canvas_tab/main/workflows/Turbo_canvas.svg)
+
+This is the basic turbo workflow.  The default view in the canvas is initialized to provide two
+images,  Image A is sent to the canvas node.  A layer on Image B is set as the target for new
+generations.
+
+You can draw and refine your work by starting with a simple scribble in Image A and when something
+you like is generated, CTRL-LeftClick on the duplicate layer button will copy that image
+and you can use that as a base drawing.
+
+![target info](https://raw.githubusercontent.com/Lerc/canvas_tab/main/workflows/TomatoDoodle.jpg)
+
+You can get into a good design process by drawing a general outline with a high denoising value, then
+successively copying the generated image and drawing on top of it for finer adjustments, gradually reducing
+the denoising value as you go.
+
+Sometimes when iterating generating images bases upon previous generations the image may start to become blurry,
+this workflow contains a node for sharpening the image that can reduce this problem,  By default the sharpening node
+is set to bypass meaning it takes no time and has no effect.   If the image becomes blurry you can reactivate this
+node for a single generation to get a sharper output to use as a new source image.
+
+
+## Multi Canvas with ControlNet
+
+![turbo workflow](https://raw.githubusercontent.com/Lerc/canvas_tab/main/workflows/TurboLLiteDepth.svg)
+This workflow uses the ControlLLite depthmap model,   It requires creating another image in the canvas editor
+with the new image button and setting the title to "Depth".  The workflow contains a canvas node with the title "Depth",  when an image matches the canvas node title the node will be updated directly from that image.
+
+https://raw.githubusercontent.com/Lerc/canvas_tab/main/workflows/MultiTarget.mp4
+
+
+
+
+