- Imagine Weekly – The only Midjourney Newsletter you need
- Posts
- Imagine Weekly – Issue 11 [August 16th]
Imagine Weekly – Issue 11 [August 16th]
Welcome to Issue 11 of Imagine Weekly 👋
Hello everyone,
This week, the MidJourney team has given us many new updated regarding the things they are working on. Besides that it’s hard to miss the hype around the FLUX model from Black Forest Labs – which indeed does generate great images. More down below. Have fun reading 📖
In this issue you’ll find:
📖 TL;DR; Weekly Recap of the Office Hours (or the extended recap)
🎓 What the FLUX?
🎨 Some new srefs to tinker with
Office Hours TL;DR; Recap 2024-08-07
🖥️ New Web Editor: Combining reframe and repaint controls, releasing today or tomorrow.
🧠 Upcoming Features: Personality survey and demographics exploration tool.
🌐 Depth Control Net: Testing with version 6.0.
🚀 Midjourney 6.2 Model: In training, testing in 1-2 weeks.
📈 Larger Batch Sizes: Testing 8 or 16 batch sizes.
🎨 Personalization System: Manage up to 10 custom models, testing next week.
📅 Release Schedule: Features releasing in the next 2-4 weeks, 6.2 targeted for end of the month.
🔮 Long-term Projects: V7 models, video models, 3D data collection, drawing features.
🌐 Website Updates: Lowered image requirement, room syncing with Discord.
📝 Miscellaneous: Changes to image rating, explore and describe features, new top bar design.
🤫 Unannounced Projects: Big secret projects to be announced by year-end.
This list could use it’s own TL;DR; as the sheer amount of news is insane. I am particularly excited about the depth control net, which will us give more control about the actual image which is being generated. Next to the new editor in the web ui looks quite nice. Later today we all should be able to tinker with it.
🚀 What the FLUX?
It's hard to miss the current hype about FLUX, a new model released by Black Forest Labs. The results are definitely worth taking a look, here are some comparison images:
Midjourney v6.1 | Flux Dev |
Midjourney v6.1: a photo of a beautiful street in Freiburg with a tram passing by and people walking about and riding bikes
Flux Example image from the website: a photo of a beautiful street in Freiburg with a tram passing by and people walking about and riding bikes
So currently I am impressed on what such a small team at Black Forest Labs achieved. I am still a big fan of MidJourney and that won’t change anytime soon. Nevertheless, especially if you run FLUX in a pipeline e.g. in ComfyUI with some other tools (tutorial will be published in the next few weeks) you can get even better results. With MidJourney that’s currently not possible. So let’s see what the future brings.
✨ Got Flux training working on photoai.com with my own photos now!!!!
It looks SO real 😱
I think we can finally fire our photographers for real now
First 2 pics from the checkpoint sampling so bit rough
Last 2 pics generated from the .safetensor trained on me
— @levelsio (@levelsio)
3:24 AM • Aug 13, 2024
Next to FLUX there’s also Grok which now supports generating images (it uses FLUX behind the scenes). This comparison by @ciguleva of MidJourney and Grok is pretty accurate I’d say. Check out her thread and reasoning:
Grok. I've reached my image generation usage limit for today 🙂
Conclusion: It's like Midjourney v5.2 but with better prompt understanding, so the old prompts I used for that MJ model work well for Grok.
— Tatiana Tsiguleva (@ciguleva)
2:12 AM • Aug 15, 2024
And again some video related content which looks and sounds quite impressive:
🚨TOKEN VIDEO SHARE🚨
I see a lot of AI video examples with people who are happy and talking. Don’t forget that many other emotions are also well handled by AI videos.Here’s a video example where I’ve highlighted the ones that work best.
Bookmark, like, and share widely for… x.com/i/web/status/1…— Pierrick Chevallier | IA (@CharaspowerAI)
5:36 PM • Aug 13, 2024
And some more FLUX examples
FLUX LoRA Resources
Mega Thread👇
— Heather Cooper (@HBCoop_)
6:12 PM • Aug 13, 2024
🎨 Inspiration, Prompts and Srefs
For today's issue, I've gathered some bold srefs for you to save and tinker with.
Moodboard is ready. Here’s the code:
--sref 2837013301 // Ancient Gold
Sample prompt in the ALT
Midjourney Styles + Personalization
— Dogan Ural (@doganuraldesign)
2:30 PM • Aug 15, 2024
NEW --sref 3190971442
Very colourful photographic style that will default to neon pink, yellow and blue. Great for fashion, product, food, lifestyle, and wildlife. If you like colours you'll enjoy this style.
Works with --sw 100 --stylize 300
Share your creations | #Midjourneyx.com/i/web/status/1…— Michael Rabone (@michaelrabone)
5:02 PM • Aug 12, 2024
See you next week 👋🤖
I hope you enjoyed this issue of Imagine Weekly. If you found it helpful, we'd be thrilled if you could share it with your colleagues and friends. Please feel free to reach out with any suggestions or feedback.
Have a great weekend.
To support me:
😘 Recommend the newsletter to your friends: it really helps!
📬 If you’re not subscribed yet to my newsletter Imagine Weekly, I’d be thrilled to welcome you on board!