Administrator Popular Post JuanP Posted March 12 Administrator Popular Post Share Posted March 12 AI Visualizer What is the AI Visualizer? The AI Visualizer, powered by Stable Diffusion, can create illustrative or realistic images based on your models and drawings directly inside Vectorworks. It uses natural language prompts or a combination of prompts and a Vectorworks drawing or model view to determine the image content. It uses Vectorworks Cloud Services for a zero-install, uninterrupted workflow. It is an experimental feature using an exciting and fast changing technology that has lots of potential within the design process. We expect additional features to be developed over time, such as predefined prompts and additional geometry recognition. What AI Visualizer is not AI Visualizer is not a rendering engine, such as Renderworks, in the sense that it does not calculate the 3-dimensional scene. Instead, it detects edges in the image supplied from within Vectorworks (whether this is from a 3D view, a viewport or simply a bitmap). As such, even with the Creativity slider set to 0, absolute replication of the intended image geometry is not guaranteed. For this reason, we encourage you to look at this first release of this technology as a tool for inspiration - whether it’s whole concept inspiration (giving you new ideas for projects/designs) or inspiration for different material use or illustrative styles. When and why should I use the AI Visualizer? This feature provides fast image generation for design exploration. It offers an alternative method for creating and illustrating design ideas. It allows you to explore more visual options in less time. It is directly available in the Vectorworks desktop app interface (Model>AI Visualizer), but makes use of Vectorworks Cloud processing, so it is fast, independent of your computer hardware and allows for non-blocking workflows, meaning you can carry on working in Vectorworks any time images are being generated. Is the AI Visualizer compatible with my version of Vectorworks? The AI Visualizer was introduced in Vectorworks 2024 Update 4 and is now accessible to all customers with active Subscriptions and Vectorworks Service Select members across all products (Fundamentals, Architect, Landmark, Spotlight). This feature is currently not available in academic licenses. What products does the AI Visualizer work with? The AI Visualizer works with all Vectorworks products: Architect, Landmark, Spotlight, and Fundamentals. What are the installation and system requirements when using AI Visualizer? The AI Visualizer comes automatically loaded upon installation of Vectorworks 2024 Update 4. It is accessed directly through your Vectorworks desktop interface (Model menu > AI Visualizer). The feature uses Vectorworks Cloud processing (available to all Subscription and Vectorworks Service Select members), placing no additional demands on your hardware. To activate it, you simply need to be signed in to your Vectorworks account. The VCS app must be updated to version 13.4.2736 or later. To update the VCS app, go to the About page and check for updates. To use the AI Visualizer, you must be signed into your VCS account and have the VCS desktop client running. Are there any getting started tips? This feature relies on text prompts from the user, and certain text prompts work better than others. The online Help includes tips and tricks on creating successful text prompts. You don’t need a model or existing image to use the AI Visualizer. Images can be generated from text prompts alone. The AI Visualizer works best with text prompts in English, but it is possible to use other languages. As it can generate images from text prompts alone, it has the potential to be used to create images for use in custom image props, texture patterns, or 360 panoramic backgrounds. Anything else I need to know? This is an image generator only; it cannot generate modelled objects, make changes to a Vectorworks model, or change render settings. Unlike other images and drawings that you create in Vectorworks, AI-generated images cannot be copyrighted. Vectorworks collects user data associated with text prompts and images, archives the data for seven days, and then deletes it. Future development Our development team is already working on additional ideas for the development of this feature. Additional capabilities such as improved geometry recognition, predefined prompts, saved prompts, and depth processing are already developments we want to explore further. 21 2 Quote Link to comment
Todd McCurdy Posted March 12 Share Posted March 12 Thanks Juan. That clarifies things! 2 Quote Link to comment
E|FA Posted March 12 Share Posted March 12 Thanks dfor the info and where to find the Model>AI Visualizer command to add manually to custom workspaces. 1 Quote Link to comment
VIRTUALENVIRONS Posted March 13 Share Posted March 13 Do you need Service Select to see AI Visualizer? Just loaded Version 4 and I don't have the Ai Visualizer option. Quote Link to comment
Administrator JuanP Posted March 13 Author Administrator Share Posted March 13 30 minutes ago, VIRTUALENVIRONS said: Do you need Service Select to see AI Visualizer? Yes, that is correct. Is the AI Visualizer compatible with my version of Vectorworks? The AI Visualizer was introduced in Vectorworks 2024 Update 4 and is now accessible to all customers with active Subscriptions and Vectorworks Service Select members across all products (Fundamentals, Architect, Landmark, Spotlight). Quote Link to comment
Tom W. Posted March 13 Share Posted March 13 48 minutes ago, VIRTUALENVIRONS said: Do you need Service Select to see AI Visualizer? Just loaded Version 4 and I don't have the Ai Visualizer option. If you're using a custom workspace you will need to add it manually following the update. Quote Link to comment
Vectorworks, Inc Employee Tamsin Slatter Posted March 13 Vectorworks, Inc Employee Share Posted March 13 You'll also need to update your Vectorworks Cloud Services app. Just open it, go to the About menu and run Refresh. Then restart Vectorworks. Quote Link to comment
VIRTUALENVIRONS Posted March 13 Share Posted March 13 Thanks all for the info, but I think I will let this pass. Looks like a fun and useful function. While I have VW attention. V4 is already problematic. I found 3.1 quite stable. Will probably reload that version if I can. I think this is a bug, but not sure. Has to do with the Home Screen. Paul Quote Link to comment
Administrator JuanP Posted March 13 Author Administrator Share Posted March 13 @VIRTUALENVIRONS Unfortunately, I can't replicate your issue. A couple of things to try: - Reboot your system - Go to the Applications folder>Vectorworks 2024> and launch the Vectorworks Updater>Advanced Options> Repair If you still have issues, please get in touch with our support team via: https://customers.vectorworks.net/support Quote Link to comment
Administrator JuanP Posted March 13 Author Administrator Share Posted March 13 @VIRTUALENVIRONS It was brought to my attention that "All Home Screen images/links for Architect, Landmark, Spotlight, Fundamentals, and Designer have been updated as of 10AM EST." Maybe that could be the cause of your issue. Quote Link to comment
VIRTUALENVIRONS Posted March 13 Share Posted March 13 I did the repair and it loaded missing files all "Darwin and Core". Did a full restart, but no change. Thanks for the help. Quote Link to comment
Vectorworks, Inc Employee Hugues Posted March 13 Vectorworks, Inc Employee Share Posted March 13 @VIRTUALENVIRONS If the Vectorworks Cloud Service app is running, sign out and restart it, or sign out from Vectorworks and sign in again. This should resolve the issue. Quote Link to comment
VIRTUALENVIRONS Posted March 13 Share Posted March 13 I have never used Cloud Service (that I know of). Not sure what you mean when you say "sign out'. Do you mean restart? Quote Link to comment
Vectorworks, Inc Employee Hugues Posted March 13 Vectorworks, Inc Employee Share Posted March 13 1 hour ago, VIRTUALENVIRONS said: I have never used Cloud Service (that I know of). Not sure what you mean when you say "sign out'. Do you mean restart? If you aren't running the Cloud Service app, sign out of Vectorworks from this button on the top right corner, then shut down and restart Vectorworks. Sign-in again and the problem should go away. Quote Link to comment
VIRTUALENVIRONS Posted March 13 Share Posted March 13 Thank you for that. I have uninstalled the newest Version. What I would really like to do is re-install Version 3.1. Is that possible? Quote Link to comment
rDesign Posted March 13 Share Posted March 13 1 hour ago, VIRTUALENVIRONS said: Thank you for that. I have uninstalled the newest Version. What I would really like to do is re-install Version 3.1. Is that possible? See instructions in the following post for how to revert / downgrade to an old Service Pack, instructions are for 2023 but are the same for 2024. Quote Link to comment
Flair-Studio Posted March 13 Share Posted March 13 This is good to see and it is great that it works pretty seamlessly, the downside for the moment is that it is still quite limited in terms of options to control the outcome. Can't wait to see what's coming for the next update! Quote Link to comment
Popular Post James Russell Posted March 14 Popular Post Share Posted March 14 Hi All, Firstly, congratulations to the Vectorworks Development team - this is a bold step and a necessary implementation to remain competitive within the market - and to simplify what some of us were turning to other applications (SD Standalone, Photoshop, etc) for in the interim. Initial implementation, in my opinion, is great. Updated VW2024.4, Updated VSC, Pulled 'AI Visualizer' into the workspace and generating images in no time. Coming from a background of using Stable Diffusion 1.4 (2022) to XL1.0 I'd like to put the following items up for discussion and implementation - primarily to bring the usability of this feature inline with longterm SD user expectations; Saving and File Names Organisation of files is one of the baseline features of SD iterations. As such the following could be considered; Option to Automatically Save All Iterations to Folder. Option to Automatically Create a Session Folder. Option to Auto Name files with; Incremental Number Datestamp / Timestamp Shorthand Description (Initial 32 / 64 Characters of Description Field) Seed Number Option to Save Upscaled Versions to Separate Folder Option for Text file to be included with; Associated Incremental Number (ID) Within text file include full details of generation; Positive Descriptions, Negative Descriptions, Seed Number, Image Resolution, Number of Steps Super important when leaving a project for a time, then returning and trying to remember prompts for a certain image / result / seed. Multiple Iterations (Batch) Combined with MultiSave (above) multiple iterations is one of the biggest creative boosts I've personally found over the few years of SD use - to be able to generate 8 / 16 / 64 versions of something and choose the best one. Multiple Iterations - allows multiple images to be generated (1 - 64) and then once a desired seed is found tweeks to prompts can be made. Batch Time Saving - I'd be quite keen to finalise a viewport / image in VW and then set a batch of 16 iterations going to give me options. Grid Return - natively SD support a grid summary of results with individual outputs to follow - suggest implementing. Iteration Slider Controls Although the creativity slider is good, and is a simplification of the multitude of variance in SD it would be good to see some of the following in an 'Advanced' dropdown; Seed Number should be shown for resuming previous sessions (presumably this is what the Generate Similar actually refers to) For new users the Creativity Slider should indicate 'More Creativity' direction (hover text same as the CFG slider native to SD for reference). Number of steps - combined with iteration steps becomes and important element in 'steering' your desired results. I'll often produce 64 results at 25 Steps, followed by the best of those 24 x 40 Steps and then the best of those 12 x 60 Steps. In Fill / In Painting Very sure this must be in the pipeline however worth mentioning. Marquee Selection Infill - a logical step to be able to select a particular element and infill. Rectangle, circle and lasso marquee and mask painting overlay would be preferred. Inverse Selection Infill - useful for keeping a central element (for example an outdoor stage) and then changing the surrounding elements (sky / weather / season) Bleed / Blur / Feathering - the 'hardness' of the marquee selection and resultant Infill / Outfill. Miscellaneous Items Face Restoration - could very well be on by default. Would advise allowing for pre- and post- images to be saved independently for comparison. Tilling - would suggest looking at this checkbox / plugin, mainly for iterative texture creation (one of the biggest things I currently use SD for) Auto Prompting - at a GUI level it would eventually be good to see this, in a similar way to Styled Prompts in SD Standalone / Photoshop. PNG Transparency - often achieved with the Background Remover Plugin, this is the best way to create Image Props (currently how I do them with SD) I'm sure that some / most / all of these above is in the pipeline and completely consider the current release as a taste of what's to come - excellent for an introduction to many in the community. Raw image - Palais Theatre (Melbourne) Concept Generation - RAW AI VW Concept Generation - RAW AI VW Concept Generation - RAW AI VW Congratulation to the team again on this. Cheers, James 9 Quote Link to comment
Vectorworks, Inc Employee Dave Donley Posted March 14 Vectorworks, Inc Employee Share Posted March 14 Thanks for that @James Russell logged your wishes for our backlog of items for the next version(s). Can you elaborate on auto-prompting a little? Quote Link to comment
Popular Post James Russell Posted March 14 Popular Post Share Posted March 14 @Dave Donley certainly! When using SD (or other AI Image Generators; Midjourney, DALL-E, etc) for extended periods of time you'll realise that although the sliders are good, it's actually the prompts (both positive and negative) which are better at 'steering' the direction of an image. I think in my mind 'Prompt Assistance and Understanding' might have been a better way of writing my dot-point, and it comes in two parts; Parameter Weighting Understanding For those of you newer to the AI world, the Prompt essentially steers your image generation (combined with the Seed Number for variance) over several iterations through the magical web of possibility until you reach your destination image. Weighting is the concept of how much impact a particular word(s) has within the phrase. I would suggest reading this nice snippet / overview - however @Dave Donley you and the team would have to confirm that all of this syntax is accepted by VW AI Viz: https://getimg.ai/guides/guide-to-stable-diffusion-prompt-weights Within the provided VW Documentation (https://app-help.vectorworks.net/2024/eng/VW2024_Guide/Rendering2/Generating_AI_images.htm?rhmapfs=true), it would seem the GUI sees () and [] as increase and decrease in prompting respectively; for those of you used to SD (in particular Automatic1111) this is the common syntax. Guide Snippet (To be confirmed with VW): single words without parentheses: penguin holding a beer+ single or multiple words with parentheses: penguin holding a (beer)+ penguin holding a (beer)- penguin (holding a beer)+ penguin (holding a beer)- more symbols result in more effect: penguin (holding a beer)+++ penguin (holding a beer)+++ nesting: penguin (holding a beer+)++ --> beer effectively gets +++ all of the above with numbers: penguin holding (a beer)1.1 penguin (holding (a beer)1.3)1.1 Weighting with numbers + is equivalent to 1.1 ; ++ is equivalent to 1.1^2 ; +++ is equivalent to 1.1^3 etc. - is equivalent to 0.9 ; -- is equivalent to 0.9^2 ; --- is equivalent to 0.9^3 etc. As you can see, with the same Seed Value, the weighting of particular word(s) [in this case beer] can change things. In a practical example, which I'll need to test within VW AI Viz, this could be the difference between; rock band on stage, bright red shiny drum kit, male lead singer with long black hair and black vest, blue lighting, brick back wall AND (rock band) on stage, (bright red) (shiny)++ drum kit, (male lead singer) with ((long) black hair) and (black vest), blue lighting, [brick back wall] The best implementation I've seen of this was a plugin (I'll go digging later) which simply coloured more positive words / phrases in deeper shades of blue, and more negative in red, which visually helped people understand the weighting so much more! Prompt Assistance (including Styles) Over time users will find series of words which help 'steer' an AI Image Generator toward their result, which could develop into a style. Without wanting to re-train and/or checkpoint (really time consuming ways of making an AI do the particular thing you want), it's often better to find your descriptive style and then vary from there. When looking for inspiration try sites like; https://lexica.art https://prompthero.com/ These sites allow you to search for something basic, for example 'stage', and then visually find something which matches your requirements - allowing you to see the word associations with those images. So for example searching for 'stage', I might find an image that I like the look of, and from there a portion of or whole prompt of; image of a retro stage, full view of the stage, red curtains on the sides, with a slight shadow,stage lights, a soft image, a winding path of strings and notes, men dancing in the middle of the stage, women dancing in dress style,dancing on stage, orchestra backstage, ((best quality) ), ((masterpiece) ), ((realistic) ),(detailed) , creative, poster style, highly detailed realism, 4k Over time I might find myself constantly using "((best quality) , ((masterpiece)), ((realistic)), (detailed), creative, poster style, highly detailed realism, 4k" and if I'm using this phrase all the time I might like to consider this a 'Style', and save it as such: 'Custom Style - Realistic' - very much like our Renderworks Styles and Sketch Styles. So instead of writing these words every time I can just add my 'Custom Style - Realistic' and these words are added to my prompt. There's been some amazing GUIs over the last few years which have prompted users who start looking for additional words, for example if a user typed the word 'Realistic', the prompter will then suggest; masterpiece, best quality, detailed, high realism, 4k, etc I think implementing Styles or Prompt Suggestions could be something to look at for the future, in particular as this technology is introduced to newer users - essentially an automatic version of the overview provided in the "Tips for writing a good prompt" section of your existing documentation. Happy generating everyone! Cheers, James 7 Quote Link to comment
YoustenGarden Posted March 15 Share Posted March 15 Dear Administrator, Yesterday, when I opened my WV Landmark 2024 I noticed that the update 4 is available, despite no VSS at that moment. I installed it as usual, however after installation and restring my computer VW program hardly responds, looks like frozen. What does this mean? What I need to do now to make my program working? Do I need to uninstall update 4? and how? And why this update appear available for me if I do not have VSS? Thank you for your information. Regards Quote Link to comment
LCarm Posted March 15 Share Posted March 15 On 3/13/2024 at 6:39 AM, Tamsin Slatter said: You'll also need to update your Vectorworks Cloud Services app. Just open it, go to the About menu and run Refresh. Then restart Vectorworks. Can you walk me through this? I don't see "refresh". Cloud services is running, and says it is up to date. I did all the steps above except for the "refresh" because I can't find it....but I also still do not see AI Visualizer in the model menu. Thank you! Quote Link to comment
Eric Marin Posted March 16 Share Posted March 16 Does this mean that the AI Visualizer isn´t available? Ia have been unable to get the command line, even doing everything that has been mentioned here Quote Link to comment
Vectorworks, Inc Employee Tamsin Slatter Posted March 18 Vectorworks, Inc Employee Share Posted March 18 On 3/15/2024 at 5:41 PM, LCarm said: Can you walk me through this? I don't see "refresh". Cloud services is running, and says it is up to date. I did all the steps above except for the "refresh" because I can't find it....but I also still do not see AI Visualizer in the model menu. Thank you! If you're up to date on the Cloud Services app, all is good with that and you won't see the Refresh button. Within Vectorworks, if you don't see it on the Model menu, you may be using a workspace that has saved itself in your user folder and is therefore considered a "custom" workspace. You can add the menu item yourself using Tools > Workspaces > Edit Current workspace. And you may also find it under the Windows > Palettes > Web palettes menu. Good luck! Quote Link to comment
Vectorworks, Inc Employee Tamsin Slatter Posted March 18 Vectorworks, Inc Employee Share Posted March 18 On 3/16/2024 at 6:57 PM, Eric Marin said: Does this mean that the AI Visualizer isn´t available? Ia have been unable to get the command line, even doing everything that has been mentioned here I'm not sure... why you are not able to see it. See my note above about adding it to the workspace. The AI visualizer is processed on Vectorworks Cloud Services, and is available only with a subscription or as a benefit of an active Vectorworks Service Select agreement. 1 Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.