From a simple landing where we tested our MVP to a platform that distributes 20,000 codes per month, we learned a lot.

Starting point

Kwotet was my first startup. Everyone might post book quotes online.

I wanted a change.

Kwotet lacked attention, thus I felt stuck. After experiencing the trials of starting Kwotet, I thought of developing a waitlist service, but I required a strong co-founder.

I knew Dries from school, but we weren't close. He was an entrepreneurial programmer who worked a lot outside school. I needed this.

We brainstormed throughout school hours. We developed features to put us first. We worked until 3 am to launch this product.

Putting in the hours is KEY when building a startup

The instant that we lost our spark

In Belgium, college seniors do their internship in their last semester.

As we both made the decision to pick a quite challenging company, little time was left for Lancero.

Eventually, we lost interest. We lost the spark…

The only logical choice was to find someone with the same spark we started with to acquire Lancero.

And we did @ MicroAcquire.

Sell before your product dies. Make sure to profit from all the gains.

What did we do following the sale?

Not far from selling Lancero I lost my dad. I was about to start a new company. It was focused on positivity. I got none left at the time.

We still didn’t let go of the dream of becoming full-time entrepreneurs. As Dries launched the amazing company Plunk, and I’m still in the discovering stages of my next journey!

Dream!

You’re an entrepreneur if:

• You're imaginative.

• You enjoy disassembling and reassembling things.

• You're adept at making new friends.

• YOU HAVE DREAMS.

You don’t need to believe me if I tell you “everything is possible”… I wouldn't believe it myself if anyone told me this 2 years ago.

Until I started doing, living my dreams.

What Is That?

Berkus is a pre-revenue valuation method based exclusively on qualitative criteria, like Scorecard.

Few firms match their financial estimates, especially in the early stages, so valuation methodologies like the Berkus method are a good way to establish a valuation when the economic measures are not reliable.

How does it work?

This technique evaluates five key success factors.

• Fundamental principle

• Technology

• Execution

• Strategic alliances in its primary market

• Production, followed by sales

The Berkus technique values the business idea and four success factors. As seen in the matrix below, each of these dimensions poses a danger to the startup's success.

It assigns $0-$500,000 to each of these beginning regions. This approach enables a maximum $2.5M pre-money valuation.

This approach relies significantly on geography and uses the US as a baseline, as it differs in every country in Europe.

A set of standards for analyzing each dimension individually

Fundamental principle (or strength of the idea)

Ideas are worthless; execution matters. Most of us can relate to seeing a new business open in our area or a startup get funded and thinking, "I had this concept years ago!" Someone did it.

The concept remains. To assess the idea's viability, we must consider several criteria.

• The concept's exclusivity It is necessary to protect a product or service's concept using patents and copyrights. Additionally, it must be capable of generating large profits.

• Planned growth and growth that goes in a specific direction have a lot of potential, therefore incorporating them into a business is really advantageous.

• The ability of a concept to grow A venture's ability to generate scalable revenue is a key factor in its emergence and continuation. A startup needs a scalable idea in order to compete successfully in the market.

• The attraction of a business idea to a broad spectrum of people is significantly influenced by the current socio-political climate. Thus, the requirement for the assumption of conformity.

• Concept Validation Ideas must go through rigorous testing with a variety of audiences in order to lower risk during the implementation phase.

Technology (Prototype)

This aspect reduces startup's technological risk. How good is the startup prototype when facing cyber threats, GDPR compliance (in Europe), tech stack replication difficulty, etc.?

Execution

Check the management team's efficacy. A potential angel investor must verify the founders' experience and track record with previous ventures. Good leadership is needed to chart a ship's course.

Strategic alliances in its primary market

Existing and new relationships will play a vital role in the development of both B2B and B2C startups. What are the startup's synergies? potential ones?

Production, followed by sales (product rollout)

Startup success depends on its manufacturing and product rollout. It depends on the overall addressable market, the startup's ability to market and sell their product, and their capacity to provide consistent, high-quality support.

Example

We're now founders of EyeCaramba, a machine vision-assisted streaming platform. My imagination always goes to poor puns when naming a startup.

Since we're first-time founders and the Berkus technique depends exclusively on qualitative methods and the evaluator's skill, we ask our angel-investor acquaintance for a pre-money appraisal of EyeCaramba.

Our friend offers us the following table:

Because we're first-time founders, our pal lowered our Execution score. He knows the idea's value and that the gaming industry is red-hot, with worse startup ideas getting funded, therefore he gave the Basic value the highest value (idea).

EyeCaramba's pre-money valuation is $400,000 + $250,000 + $75,000 + $275,000 + $164,000 (1.16M). Good.

References

• https://medium.com/humble-ventures/how-angel-investors-value-pre-revenue-startups-part-iii-8271405f0774#:~:text=pre%2Drevenue%20startups.-,Berkus%20Method,potential%20of%20the%20idea%20itself.%E2%80%9D

• https://eqvista.com/berkus-valuation-method-for-startups/

• https://www.venionaire.com/early-stage-startup-valuation-part-2-the-berkus-method/

Scaling your startup requires answering customer demands, not growth tricks.

I cofounded Freedo Rentals in 2019. I reached 50 lakh+ ARR in 6 months before quitting owing to the epidemic.

Freedo aimed to solve 2 customer pain points:

• Users lacked a reliable last-mile transportation option.

• The amount that Auto walas charge for unmetered services

Solution?

Effectively simple.

Build ports at high-demand spots (colleges, residential societies, metros). Electric ride-sharing can meet demand.

We had many problems scaling. I'll explain using the AARRR model.

• Brand unfamiliarity or a novel product offering were the problems with awareness. Nobody knew what Freedo was or what it did.

• Problem with awareness: Content and advertisements did a poor job of communicating the task at hand. The advertisements clashed with the white-collar part because they were too cheesy.

• Retention Issue: We encountered issues, indicating that the product was insufficient. Problems with keyless entry, creating bills, stealing helmets, etc.

• Retention/Revenue Issue: Costly compared to established rivals. Shared cars were 1/3 of our cost.

• Referral Issue: Missing the opportunity to seize the AHA moment. After the ride, nobody remembered us.

Once you know where you're struggling with AARRR, iterative solutions are usually best.

Once you have nailed the AARRR model, most startups use paid channels to scale. This dependence, on paid channels, increases with scale unless you crack your organic/inbound game.

Over-index growth loops. Growth loops increase inflow and customers as you scale.

When considering growth, ask yourself:

• Who is the solution's ICP (Ideal Customer Profile)? (To whom are you selling)

• What are the most important messages I should convey to customers? (This is an A/B test.)

• Which marketing channels ought I prioritize? (Conduct analysis based on the startup's maturity/stage.)

• Choose the important metrics to monitor for your AARRR funnel (not all metrics are equal)

• Identify the Flywheel effect's growth loops (inertia matters)

My biggest mistakes:

• not paying attention to consumer comments or satisfaction. It is the main cause of problems with referrals, retention, and acquisition for startups. Beyond your NPS, you should consider second-order consequences.

• The tasks at hand should be quite clear.

Here's my scaling equation:

Growth = A x B x C

A = Funnel top (Traffic)

B = Product Valuation (Solving a real pain point)

C = Aha! (Emotional response)

Freedo's A, B, and C created a unique offering.

Freedo’s ABC:

A — Working or Studying population in NCR

B — Electric Vehicles provide last-mile mobility as a clean and affordable solution

C — One click booking with a no-noise scooter

Final outcome:

FWe scaled Freedo to Rs. 50 lakh MRR and were growing 60% month on month till the pandemic ceased our growth story.

How we did it?

We tried ambassadors and coupons. WhatsApp was our most successful A/B test.

We grew widespread adoption through college and society WhatsApp groups. We requested users for referrals in community groups.

What worked for us won't work for others. This scale underwent many revisions.

Every firm is different, thus you must know your customers. Needs to determine which channel to prioritize and when.

Users desired a safe, time-bound means to get there.

This (not mine) growth framework helped me a lot. You should follow suit.

If you think NFTs are awful, check out the art market.

The fervor around NFTs has subsided in recent months due to the crypto market crash and the media's short attention span. They were all anyone could talk about earlier this spring. Last semester, when passions were high and field luminaries were discussing "slurp juices," I asked my students and students from over 20 other universities what they thought of NFTs.

According to many, NFTs were either tasteless pyramid schemes or a new way for artists to make money. NFTs contributed to the climate crisis and harmed the environment, but so did air travel, fast fashion, and smartphones. Some students complained that NFTs were cheap, tasteless, algorithmically generated schlock, but others asked how this was different from other art.

I'm not sure what I expected, but the intensity of students' reactions surprised me. They had strong, emotional opinions about a technology I'd always considered administrative. NFTs address ownership and accounting, like most crypto/blockchain projects.

Art markets can be irrational, arbitrary, and subject to the same scams and schemes as any market. And maybe a few shenanigans that are unique to the art world.

The Fairness Question

Fairness, a deflating moral currency, was the general sentiment (the less of it in circulation, the more ardently we clamor for it.) These students, almost all of whom are artists, complained to the mismatch between the quality of the work in some notable NFT collections and the excessive amounts these items were fetching on the market. They can sketch a Bored Ape or Lazy Lion in their sleep. Why should they buy ramen with school loans while certain swindlers get rich?

I understand students. Art markets are unjust. They can be irrational, arbitrary, and governed by chance and circumstance, like any market. And art-world shenanigans.

Almost every mainstream critique leveled against NFTs applies just as easily to art markets

Over 50% of artworks in circulation are fake, say experts. Sincere art collectors and institutions are upset by the prevalence of fake goods on the market. Not everyone. Wealthy people and companies use art as investments. They can use cultural institutions like museums and galleries to increase the value of inherited art collections. People sometimes buy artworks and use family ties or connections to museums or other cultural taste-makers to hype the work in their collection, driving up the price and allowing them to sell for a profit. Money launderers can disguise capital flows by using market whims, hype, and fluctuating asset prices.

Almost every mainstream critique leveled against NFTs applies just as easily to art markets.

Art has always been this way. Edward Kienholz's 1989 print series satirized art markets. He stamped 395 identical pieces of paper from $1 to $395. Each piece was initially priced as indicated. Kienholz was joking about a strange feature of art markets: once the last print in a series sells for $395, all previous works are worth at least that much. The entire series is valued at its highest auction price. I don't know what a Kienholz print sells for today (inquire with the gallery), but it's more than $395.

I love Lee Lozano's 1969 "Real Money Piece." Lozano put cash in various denominations in a jar in her apartment and gave it to visitors. She wrote, "Offer guests coffee, diet pepsi, bourbon, half-and-half, ice water, grass, and money." "Offer real money as candy."

Lee Lozano kept track of who she gave money to, how much they took, if any, and how they reacted to the offer of free money without explanation. Diverse reactions. Some found it funny, others found it strange, and others didn't care. Lozano rarely says:

Apr 17 Keith Sonnier refused, later screws lid very tightly back on. Apr 27 Kaltenbach takes all the money out of the jar when I offer it, examines all the money & puts it all back in jar. Says he doesn’t need money now. Apr 28 David Parson refused, laughing. May 1 Warren C. Ingersoll refused. He got very upset about my “attitude towards money.” May 4 Keith Sonnier refused, but said he would take money if he needed it which he might in the near future. May 7 Dick Anderson barely glances at the money when I stick it under his nose and says “Oh no thanks, I intend to earn it on my own.” May 8 Billy Bryant Copley didn’t take any but then it was sort of spoiled because I had told him about this piece on the phone & he had time to think about it he said.

Smart Contracts (smart as in fair, not smart as in Blockchain)

Cornell University's Cheryl Finley has done a lot of research on secondary art markets. I first learned about her research when I met her at the University of Florida's Harn Museum, where she spoke about smart contracts (smart as in fair, not smart as in Blockchain) and new protocols that could help artists who are often left out of the economic benefits of their own work, including women and women of color.

Her talk included findings from her ArtNet op-ed with Lauren van Haaften-Schick, Christian Reeder, and Amy Whitaker.

NFTs allow us to think about and hack on formal contractual relationships outside a system of laws that is currently not set up to service our community.

The ArtNet article The Recent Sale of Amy Sherald's ‘Welfare Queen' Symbolizes the Urgent Need for Resale Royalties and Economic Equity for Artists discussed Sherald's 2012 portrait of a regal woman in a purple dress wearing a sparkling crown and elegant set of pearls against a vibrant red background.

Amy Sherald sold "Welfare Queen" to Princeton professor Imani Perry. Sherald agreed to a payment plan to accommodate Perry's budget.

Amy Sherald rose to fame for her 2016 portrait of Michelle Obama and her full-length portrait of Breonna Taylor, one of the most famous works of the past decade.

As is common, Sherald's rising star drove up the price of her earlier works. Perry's "Welfare Queen" sold for $3.9 million in 2021.

Imani Perry's early investment paid off big-time. Amy Sherald, whose work directly increased the painting's value and who was on an artist's shoestring budget when she agreed to sell "Welfare Queen" in 2012, did not see any of the 2021 auction money. Perry and the auction house got that money.

Sherald sold her Breonna Taylor portrait to the Smithsonian and Louisville's Speed Art Museum to fund a $1 million scholarship. This is a great example of what an artist can do for the community if they can amass wealth through their work.

NFTs haven't solved all of the art market's problems — fakes, money laundering, market manipulation — but they didn't create them. Blockchain and NFTs are credited with making these issues more transparent. More ideas emerge daily about what a smart contract should do for artists.

NFTs are a copyright solution. They allow us to hack formal contractual relationships outside a law system that doesn't serve our community.

Amy Sherald shows the good smart contracts can do (as in, well-considered, self-determined contracts, not necessarily blockchain contracts.) Giving back to our community, deciding where and how our work can be sold or displayed, and ensuring artists share in the equity of our work and the economy our labor creates.

Steve Jobs' initial mission at Apple in 1997 was to destroy. He killed the Newton PDA and Macintosh clones. Apple stopped trying to please everyone under Jobs.

Afterward, there were few highly targeted moves. First, the pink iMac. Modest success. The iPod, iPhone, and iPad made Apple the world's most valuable firm. Each maneuver changed the company's center of gravity and won.

That's the idea behind Schwerpunkt, a German military term meaning "focus." Jobs didn't need to win everywhere, just where it mattered, so he focused Apple's resources on a few key goods. Finding your Schwerpunkt is more important than charts and analysis for excellent strategy.

Comparison of Relative Strength and Relative Weakness

The iPod, Apple's first major hit after Jobs' return, didn't damage Microsoft and the PC, but instead focused Apple's emphasis on a fledgling, fragmented market that generated "sucky" products. Apple couldn't have taken on the computer titans at this stage, yet it beat them.

The move into music players used Apple's particular capabilities, especially its ability to build simple, easy-to-use interfaces. Jobs' charisma and stature, along his understanding of intellectual property rights from Pixar, helped him build up iTunes store, which was a quagmire at the time.

In Good Strategy | Bad Strategy, management researcher Richard Rumelt argues that good strategy uses relative strength to counter relative weakness. To discover your main point, determine your abilities and where to effectively use them.

Steve Jobs did that at Apple. Microsoft and Dell, who controlled the computer sector at the time, couldn't enter the music player business. Both sought to produce iPod competitors but failed. Apple's iPod was nobody else's focus.

Finding The Center of Attention

In a military engagement, leaders decide where to focus their efforts by assessing commanders intent, the situation on the ground, the topography, and the enemy's posture on that terrain. Officers spend their careers learning about schwerpunkt.

Business executives must assess internal strengths including personnel, technology, and information, market context, competitive environment, and external partner ecosystems. Steve Jobs was a master at analyzing forces when he returned to Apple.

He believed Apple could integrate technology and design for the iPod and that the digital music player industry sucked. By analyzing competitors' products, he was convinced he could produce a smash by putting 1000 tunes in my pocket.

The only difficulty was there wasn't the necessary technology. External ecosystems were needed. On a trip to Japan to meet with suppliers, a Toshiba engineer claimed the company had produced a tiny memory drive approximately the size of a silver dollar.

Jobs knew the memory drive was his focus. He wrote a $10 million cheque and acquired exclusive technical rights. For a time, none of his competitors would be able to recreate his iPod with the 1000 songs in my pocket.

How to Enter the OODA Loop

John Boyd invented the OODA loop as a pilot to better his own decision-making. First OBSERVE your surroundings, then ORIENT that information using previous knowledge and experiences. Then you DECIDE and ACT, which changes the circumstance you must observe, orient, decide, and act on.

Steve Jobs used the OODA loop to decide to give Toshiba $10 million for a technology it had no use for. He compared the new information with earlier observations about the digital music market.

Then something much more interesting happened. The iPod was an instant hit, changing competition. Other computer businesses that competed in laptops, desktops, and servers created digital music players. Microsoft's Zune came out in 2006, Dell's Digital Jukebox in 2004. Both flopped.

By then, Apple was poised to unveil the iPhone, which would cause its competitors to Observe, Orient, Decide, and Act. Boyd named this OODA Loop infiltration. They couldn't gain the initiative by constantly reacting to Apple.

Microsoft and Dell were titans back then, but it's hard to recall. Apple went from near bankruptcy to crushing its competition via Schwerpunkt.

Rather than a destination, it is a journey

Trying to win everywhere is a strategic blunder. Win significant fights, not trivial skirmishes. Identifying a focal point to direct resources and efforts is the essence of Schwerpunkt.

When Steve Jobs returned to Apple, PC firms were competing, but he focused on digital music players, and the iPod made Apple a player. He launched the iPhone when his competitors were still reacting. When Steve Jobs said, "One more thing," at the end of a product presentation, he had a new focus.

Schwerpunkt isn't static; it's dynamic. Jobs' ability to observe, refocus, and modify the competitive backdrop allowed Apple to innovate consistently. His strategy was tailored to Apple's capabilities, customers, and ecosystem. Microsoft or Dell, better suited for the enterprise sector, couldn't succeed with a comparable approach.

There is no optimal strategy, only ones suited to a given environment, when relative strength might be used against relative weakness. Discovering the center of gravity where you can break through is more of a journey than a destination; it will become evident after you reach.

Deep learning-based generative art is being researched. As usual, self-learning is better. Some models, like OpenAI's DALL-E 2, require registration and can only be used online, but others can be used locally, which is usually more enjoyable for curious users. I'll demonstrate the Stable Diffusion model's operation on a standard PC.

Let’s get started.

What It Does

Stable Diffusion uses numerous components:

• A generative model trained to produce images is called a diffusion model. The model is incrementally improving the starting data, which is only random noise. The model has an image, and while it is being trained, the reversed process is being used to add noise to the image. Being able to reverse this procedure and create images from noise is where the true magic is (more details and samples can be found in the paper).

• An internal compressed representation of a latent diffusion model, which may be altered to produce the desired images, is used (more details can be found in the paper). The capacity to fine-tune the generation process is essential because producing pictures at random is not very attractive (as we can see, for instance, in Generative Adversarial Networks).

• A neural network model called CLIP (Contrastive Language-Image Pre-training) is used to translate natural language prompts into vector representations. This model, which was trained on 400,000,000 image-text pairs, enables the transformation of a text prompt into a latent space for the diffusion model in the scenario of stable diffusion (more details in that paper).

This figure shows all data flow:

The weights file size for Stable Diffusion model v1 is 4 GB and v2 is 5 GB, making the model quite huge. The v1 model was trained on 256x256 and 512x512 LAION-5B pictures on a 4,000 GPU cluster using over 150.000 NVIDIA A100 GPU hours. The open-source pre-trained model is helpful for us. And we will.

Install

Before utilizing the Python sources for Stable Diffusion v1 on GitHub, we must install Miniconda (assuming Git and Python are already installed):

wget https://repo.anaconda.com/miniconda/Miniconda3-py39_4.12.0-Linux-x86_64.sh
chmod +x Miniconda3-py39_4.12.0-Linux-x86_64.sh
./Miniconda3-py39_4.12.0-Linux-x86_64.sh
conda update -n base -c defaults conda

Install the source and prepare the environment:

git clone https://github.com/CompVis/stable-diffusion
cd stable-diffusion
conda env create -f environment.yaml
conda activate ldm
pip3 install transformers --upgrade

Download the pre-trained model weights next. HiggingFace has the newest checkpoint sd-v14.ckpt (a download is free but registration is required). Put the file in the project folder and have fun:

python3 scripts/txt2img.py --prompt "hello world" --plms --ckpt sd-v1-4.ckpt --skip_grid --n_samples 1

Almost. The installation is complete for happy users of current GPUs with 12 GB or more VRAM. RuntimeError: CUDA out of memory will occur otherwise. Two solutions exist.

Running the optimized version

Try optimizing first. After cloning the repository and enabling the environment (as previously), we can run the command:

python3 optimizedSD/optimized_txt2img.py --prompt "hello world" --ckpt sd-v1-4.ckpt --skip_grid --n_samples 1

Stable Diffusion worked on my visual card with 8 GB RAM (alas, I did not behave well enough to get NVIDIA A100 for Christmas, so 8 GB GPU is the maximum I have;).

Running Stable Diffusion without GPU

If the GPU does not have enough RAM or is not CUDA-compatible, running the code on a CPU will be 20x slower but better than nothing. This unauthorized CPU-only branch from GitHub is easiest to obtain. We may easily edit the source code to use the latest version. It's strange that a pull request for that was made six months ago and still hasn't been approved, as the changes are simple. Readers can finish in 5 minutes:

• Replace if attr.device!= torch.device(cuda) with if attr.device!= torch.device(cuda) and torch.cuda.is available at line 20 of ldm/models/diffusion/ddim.py ().

• Replace if attr.device!= torch.device(cuda) with if attr.device!= torch.device(cuda) and torch.cuda.is available in line 20 of ldm/models/diffusion/plms.py ().

• Replace device=cuda in lines 38, 55, 83, and 142 of ldm/modules/encoders/modules.py with device=cuda if torch.cuda.is available(), otherwise cpu.

• Replace model.cuda() in scripts/txt2img.py line 28 and scripts/img2img.py line 43 with if torch.cuda.is available(): model.cuda ().

Run the script again.

Testing

Test the model. Text-to-image is the first choice. Test the command line example again:

python3 scripts/txt2img.py --prompt "hello world" --plms --ckpt sd-v1-4.ckpt --skip_grid --n_samples 1

The slow generation takes 10 seconds on a GPU and 10 minutes on a CPU. Final image:

Hello world is dull and abstract. Try a brush-wielding hamster. Why? Because we can, and it's not as insane as Napoleon's cat. Another image:

Generating an image from a text prompt and another image is interesting. I made this picture in two minutes using the image editor (sorry, drawing wasn't my strong suit):

I can create an image from this drawing:

python3 scripts/img2img.py --prompt "A bird is sitting on a tree branch" --ckpt sd-v1-4.ckpt --init-img bird.png --strength 0.8

It was far better than my initial drawing:

I hope readers understand and experiment.

Stable Diffusion UI

Developers love the command line, but regular users may struggle. Stable Diffusion UI projects simplify image generation and installation. Simple usage:

• Unpack the ZIP after downloading it from https://github.com/cmdr2/stable-diffusion-ui/releases. Linux and Windows are compatible with Stable Diffusion UI (sorry for Mac users, but those machines are not well-suitable for heavy machine learning tasks anyway;).

• Start the script.

Done. The web browser UI makes configuring various Stable Diffusion features (upscaling, filtering, etc.) easy:

V2.1 of Stable Diffusion

I noticed the notification about releasing version 2.1 while writing this essay, and it was intriguing to test it. First, compare version 2 to version 1:

• alternative text encoding. The Contrastive LanguageImage Pre-training (CLIP) deep learning model, which was trained on a significant number of text-image pairs, is used in Stable Diffusion 1. The open-source CLIP implementation used in Stable Diffusion 2 is called OpenCLIP. It is difficult to determine whether there have been any technical advancements or if legal concerns were the main focus. However, because the training datasets for the two text encoders were different, the output results from V1 and V2 will differ for the identical text prompts.

• a new depth model that may be used to the output of image-to-image generation.

• a revolutionary upscaling technique that can quadruple the resolution of an image.

• Generally higher resolution Stable Diffusion 2 has the ability to produce both 512x512 and 768x768 pictures.

The Hugging Face website offers a free online demo of Stable Diffusion 2.1 for code testing. The process is the same as for version 1.4. Download a fresh version and activate the environment:

conda deactivate  
conda env remove -n ldm  # Use this if version 1 was previously installed
git clone https://github.com/Stability-AI/stablediffusion
cd stablediffusion
conda env create -f environment.yaml
conda activate ldm

Hugging Face offers a new weights ckpt file.

The Out of memory error prevented me from running this version on my 8 GB GPU. Version 2.1 fails on CPUs with the slow conv2d cpu not implemented for Half error (according to this GitHub issue, the CPU support for this algorithm and data type will not be added). The model can be modified from half to full precision (float16 instead of float32), however it doesn't make sense since v1 runs up to 10 minutes on the CPU and v2.1 should be much slower. The online demo results are visible. The same hamster painting with a brush prompt yielded this result:

It looks different from v1, but it functions and has a higher resolution.

The superresolution.py script can run the 4x Stable Diffusion upscaler locally (the x4-upscaler-ema.ckpt weights file should be in the same folder):

python3 scripts/gradio/superresolution.py configs/stable-diffusion/x4-upscaling.yaml x4-upscaler-ema.ckpt

This code allows the web browser UI to select the image to upscale:

The copy-paste strategy may explain why the upscaler needs a text prompt (and the Hugging Face code snippet does not have any text input as well). I got a GPU out of memory error again, although CUDA can be disabled like v1. However, processing an image for more than two hours is unlikely:

Stable Diffusion Limitations

When we use the model, it's fun to see what it can and can't do. Generative models produce abstract visuals but not photorealistic ones. This fundamentally limits The generative neural network was trained on text and image pairs, but humans have a lot of background knowledge about the world. The neural network model knows nothing. If someone asks me to draw a Chinese text, I can draw something that looks like Chinese but is actually gibberish because I never learnt it. Generative AI does too! Humans can learn new languages, but the Stable Diffusion AI model includes only language and image decoder brain components. For instance, the Stable Diffusion model will pull NO WAR banner-bearers like this:

V1:

V2.1:

The shot shows text, although the model never learned to read or write. The model's string tokenizer automatically converts letters to lowercase before generating the image, so typing NO WAR banner or no war banner is the same.

I can also ask the model to draw a gorgeous woman:

V1:

V2.1:

The first image is gorgeous but physically incorrect. A second one is better, although it has an Uncanny valley feel. BTW, v2 has a lifehack to add a negative prompt and define what we don't want on the image. Readers might try adding horrible anatomy to the gorgeous woman request.

If we ask for a cartoon attractive woman, the results are nice, but accuracy doesn't matter:

V1:

V2.1:

Another example: I ordered a model to sketch a mouse, which looks beautiful but has too many legs, ears, and fingers:

V1:

V2.1: improved but not perfect.

V1 produces a fun cartoon flying mouse if I want something more abstract:

I tried multiple times with V2.1 but only received this:

The image is OK, but the first version is closer to the request.

Stable Diffusion struggles to draw letters, fingers, etc. However, abstract images yield interesting outcomes. A rural landscape with a modern metropolis in the background turned out well:

V1:

V2.1:

Generative models help make paintings too (at least, abstract ones). I searched Google Image Search for modern art painting to see works by real artists, and this was the first image:

I typed "abstract oil painting of people dancing" and got this:

V1:

V2.1:

It's a different style, but I don't think the AI-generated graphics are worse than the human-drawn ones.

The AI model cannot think like humans. It thinks nothing. A stable diffusion model is a billion-parameter matrix trained on millions of text-image pairs. I input "robot is creating a picture with a pen" to create an image for this post. Humans understand requests immediately. I tried Stable Diffusion multiple times and got this:

This great artwork has a pen, robot, and sketch, however it was not asked. Maybe it was because the tokenizer deleted is and a words from a statement, but I tried other requests such robot painting picture with pen without success. It's harder to prompt a model than a person.

I hope Stable Diffusion's general effects are evident. Despite its limitations, it can produce beautiful photographs in some settings. Readers who want to use Stable Diffusion results should be warned. Source code examination demonstrates that Stable Diffusion images feature a concealed watermark (text StableDiffusionV1 and SDV2) encoded using the invisible-watermark Python package. It's not a secret, because the official Stable Diffusion repository's test watermark.py file contains a decoding snippet. The put watermark line in the txt2img.py source code can be removed if desired. I didn't discover this watermark on photographs made by the online Hugging Face demo. Maybe I did something incorrectly (but maybe they are just not using the txt2img script on their backend at all).

Conclusion

The Stable Diffusion model was fascinating. As I mentioned before, trying something yourself is always better than taking someone else's word, so I encourage readers to do the same (including this article as well;).

Is Generative AI a game-changer? My humble experience tells me:

• I think that place has a lot of potential. For designers and artists, generative AI can be a truly useful and innovative tool. Unfortunately, it can also pose a threat to some of them since if users can enter a text field to obtain a picture or a website logo in a matter of clicks, why would they pay more to a different party? Is it possible right now? unquestionably not yet. Images still have a very poor quality and are erroneous in minute details. And after viewing the image of the stunning woman above, models and fashion photographers may also unwind because it is highly unlikely that AI will replace them in the upcoming years.

• Today, generative AI is still in its infancy. Even 768x768 images are considered to be of a high resolution when using neural networks, which are computationally highly expensive. There isn't an AI model that can generate high-resolution photographs natively without upscaling or other methods, at least not as of the time this article was written, but it will happen eventually.

• It is still a challenge to accurately represent knowledge in neural networks (information like how many legs a cat has or the year Napoleon was born). Consequently, AI models struggle to create photorealistic photos, at least where little details are important (on the other side, when I searched Google for modern art paintings, the results are often even worse;).

• When compared to the carefully chosen images from official web pages or YouTube reviews, the average output quality of a Stable Diffusion generation process is actually less attractive because to its high degree of randomness. When using the same technique on their own, consumers will theoretically only view those images as 1% of the results.

Anyway, it's exciting to witness this area's advancement, especially because the project is open source. Google's Imagen and DALL-E 2 can also produce remarkable findings. It will be interesting to see how they progress.