Unlock the Power of Firefly: Ask the Experts Session! 🔥

[Preran]

Hello people!

Did you have questions about Firefly or Generative AI for which you needed a trusted expert to weigh in?

Whether you’re a beginner trying to understand the basics or an advanced user looking to delve deeper into the functionality, our Ask the Experts session is the perfect opportunity to get your questions answered! 

How to Ask Your Questions 

Participating is easy! Just reply to this post with any questions about Firefly and its use in generative AI. Here are some ideas to get you started: 

• Image Quality: How can I enhance the resolution and overall quality of images created with Firefly? 
• Bias Mitigation: What are the best practices for preventing racial, gender, or age biases in AI-generated images? 
• Creative Integration: How do I maintain a balance between the AI's suggestions and my own artistic inputs when designing? 

Session Timeline 

• Ask Your Questions Now: This session is open for seven days. Our experts will be responding to your questions during this time. 
• Deadline for Questions: Post your questions by June 1, 2024, 12 AM UTC 

Drop your questions below, and let’s get this conversation started!

Looking forward to seeing what you all come up with. 

Cheers,

Preran

[KShinabery212]

I am creating content based on Firefly during my live streams on Behance for the Adobe Education Leaders Program and for YouTube/LinkedIn.
I just covered the Structure function (very cool).
Is there an Adobe Firefly Ambassador's program like the Adobe Express Ambassador's Program?
If so would love to join that as well as any beta things you have going on.

[Vivek12]

@KShinabery212  Thank you for your interest in the Adobe Firefly Ambassador Program. Unfortunately, we do not currently have such a program available. However, if we decide to launch a similar program in the future, we will surely contact you.

^VS

[@mj]

Hi,

Please consider flagging words or phrases that the gen engine may deem inappropriate. That way users can tweak their text prompts.

Also, since this is Ai, why can't FF suggest alternative prompts for users to try when prompts are deemed to fall foul of the engine's rules?

Tx

mj

iMSD

[Vivek12]

Thank you for the valuable feedback. We will share it with our product team. Meanwhile, you may try using prompt suggestions to auto-complete your prompts

[Jose Panadero]

I have a question regarding color space. As a VFX artist, it would be great if you can tell firefly to work in the 32bpc. Photoshop supports ACES (widely used in postproduction environments) but Firefly not generating images in deep color space. It is something that Adobe will add in the future?

[Kevin Stohlmeyer]

@Jose Panadero it wont even let you generate an image in PS Beta in 32 or 16 bit. 😕

[Daniel Kramer]

Hi Preran,

I would like to see Firefly Generate layouts in Express based on input, for example 'instagram post fathers day'

[droopydog500]

Hello @Daniel Kramer,

Have you tried text to template in Adobe Express?

Example:

[Sean McCormack]

Hey Preran, 

Another question. Will we ever be able to start Text to Image with the options for style etc checked before we begin. It's a wasted round each time you start with only the text to image bar. 

Thanks, 

Sean

[Kevin Stohlmeyer]

@Sean McCormack clicking "generate" in the welcome screen does not use a credit unless you input text first. Otherwise it takes you to the interface without a charge.

[Sean McCormack]

That's great to know @Kevin Stohlmeyer, thank you. 

[Tracy.H]

Hi Sean,

We've heard users complain about this and are still exploring the options with our design team. For now, you can click on any of the inspiration images from the gallery to get into the Text to Image canvas without using a generative credit.

Thanks

Tracy

[DeanUtian]

@Tracy.H - Good to hear Adobe is looking at a way of entering the Text to Image interface without having to load an example first. If you load an insprational image, you then have to clear the prompt, tags etc to start. It is a better user experience to start something from scratch.

[Chris Maddrell]

Hello Preran, I would love to see the "exclude" function restored to omit some of the persistent annoyances, eg mountains, beards, union jacks etc

[Tracy.H]

Hi Chris,

We are working on bringing this feature back and working with our models.

Thanks

Tracy

[Chris Maddrell]

I'd also like generative fill to produce results on a transparent layer so it can be easily repositioned rather than having to mask and generate a layer to blend. To get the best out of FF2 sometimes I'm working with 5 gens as overlays and their final position isn't known at the time of generation. Thanks 

[Vivek12]

Thank you for the valuable feedback. We will share it with our product team, and they will review the request.

[Chardane Edda Martinez]

This is great!

Hi Preran,

I would like to as if its possible to have a suggestive propmt ideas when generating ideas, like auto suggestion or something?

Thank you 🙂

[Kevin Stohlmeyer]

@Chardane Edda Martinez do you mean beyond the Suggestions feature? I would like to see this made available in all available languages - if you can write in French, there should be suggested prompts in French.

[Chardane Edda Martinez]

Hi Kevin, thanks for your response this is such a great help! ☺️☺️ Unfortunately, I don't know how to speak in French but this is a great help for those like me who's having a hard time to give a semi-detailed prompt 👏👍

Thanks so much for your assistance ☺️

[Kevin Stohlmeyer]

@Chardane Edda Martinez its currently available in English-only.

[Behiç Alp Aytekin]

Hi, Preran! Thank you so much for the opportunity

Do we have a chance to use and enlarge very big canvas options such as 400x300mm which images created with Firefly?

[Tracy.H]

Hi Behiç Alp Aytekin

Our team is currently working on training our models with higher resolution for the future. The plans for what we offer our customers is still up for discussion.

Thanks

Tracy

[Chris Maddrell]

Another thought, maybe have a hover box over each effect with a short explanation of what each aims to achieve? I don't try many of them because of my finite number of credits.

[droopydog500]

Hi @Chris Maddrell,

it would be great to have a guide detailing each style. This article describes several styles: Explore art styles for AI prompts. That is the most detail I have been able to find.

I also discovered that the developer documentation for the API had larger example images for each style which makes it a little easier to study the thumbnails: Image Model Styles (API) 

    droopy