Midjourney 5.1 Parameters
The featured image for this post and the images coming up were created with natural language prompts on Midjourney (AI Image Generation Tool). A quick introduction to Midjourey in my experience is ..
It is a state-of-art tool for image generation. If any time in the past, while using an image generation AI you felt disheartened, dissatisfied then Midjourney will surprise you with the quality of end results. The awe factor that one is used to with Large Language Models like GPT, has set the benchmark high for any AI deployment. However, Midjourney delivers the same impact, visually, literally.
Regular natural language prompts work, however I found the following parameters useful. Try them, if you must.
--ar 16:9 (--ar stands for aspect ratio)--v 5.1 (--v is the version of Midjourney one wants to use. Default recently has been upgraded to 5.1, as of 3rd May, 2023, I think🤔)--s (--s is styling, it incorporates **Midjourney's** natural styling to the specified degree. Options range from 0-1000)--chaos (--chaos stands for randomisation of results. Ranges from 0 - 100)
Interestingly to use Midjourney one needs a bit of familiarisation with Discord. It is gaming/activity VoIP & Chat platform. The signup is free, interface is modern and slick. Usage is relatively easy with superb documents and tons of tutorials online. Adoption time in total is less than 30 minutes for Discord, by my estimate for a new user.
Let us now look at the images and respective prompts, these were generated in the first hour of my first interaction with Midjourney. I am sure you can tell by the excitement you’ve sensed so far.
a burger in the center of a white marble table, with ketch up and sauces painted on the table in gourmet style around the burger. fun set to stunning, intricately detailed, beautifully color-coded, octane render, hyper realistic, 8k --v 5
Girls in a pub drinking vodka shots hyper realistic, 8K --ar 16:9 --v 5 --s 500
Diver in the ocean surronded by schools of manta rays, eagle rays and barracudas **hyper realistic, 8K**
Whale in the solar system hyper detailed, hyper realistic, 8K --ar 16:9 --v 5.1 --s 500
F**our soliders walking back from a kill, with choppers and sunset in the background stunning, intricately detailed, beautifully color-coded, octane render, hyper realistic, 8k --v 5.1**
Clearly the world is changing very rapidly. Especially for knowledge workers. A new era, unlike any before may be at the horizon for the doers. Ones with an end-to-end vision working with smart people super enhanced by modern AI. Enhancement is not about surpassing, but about efficiency. As a specie we can now produce more written information and visual media with an unprecedented degree of control and access to express.
How we shape these tools, and ourselves through them is what needs focus. So use the tools, create good !
Keep reading
When Intelligence Becomes a Utility, Do You Still Need to Own the Model?
ASTRA 6.1 at roughly one-fifth the price makes me question one of the most repeated arguments in enterprise AI today: that companies should own their models. If intelligence becomes a utility, the moat may never have been the model.
Models That Choose, Not Generate: Why the Decision Layer Is Becoming Its Own Category
TypeSafe AI released Jev on 15 September. Within a fortnight there were open alternatives. The interesting part is not any single model — it is that the industry has started treating 'decide' as a different problem from 'generate', and building separate infrastructure for it.
Laya-MLX as a Sub-15 ms Decision Layer for Orchestrating Specialised SLMs on Apple Silicon
Multi-model systems built from specialised Small Language Models need something that decides, for every request, which model should answer. Using a generative model for that spends latency, tokens and parsing effort on what is fundamentally a classification problem. A research paper on typed decisions as a routing primitive.