This report from Mark Gurman has stuck with me:

In the News and Stocks apps, the display ads are no different than what you might get on an ad-supported website. In the App Store, the ads are for actual apps, which are probably more useful for Apple users than mortgage rates. Some people may resent Apple putting ads in the News and Stocks apps. After all, the iPhone is supposed to be a premium device. Let’s say you shelled out $1,000 or more to buy one, do you want to feel like Apple is squeezing more money out of you just to use its standard features? Now, a portion of ad revenue from the News app’s Today tab goes to publishers, but it’s not clear how much. Apple also lets publishers advertise within their stories and keep the vast majority of that money. Surprisingly, Today ads also appear if you subscribe to News+ for $10 per month (though it’s a smaller number).

I use Apple News often. It's a good general news catch-up tool, like Twitter without the BS. Customized notifications are helpful. Fast and lovely. Except for advertisements. I have Apple One, which includes News+, and while I understand why the magazines still have brand ads, it's ridiculous to me that Apple enables web publishers to introduce awful ads into this experience. Apple's junky commercials are ridiculous.

We know publishers want and probably requested this. Let's keep Apple News ad-free for the much smaller percentage of paid users, and here's your portion. (Same with Stocks, which is more sillier.)

Paid app placement in the App Store is a wonderful approach for developers to find new users (though far too many of those ads are trying to trick users, in my opinion).

Apple is also planning to increase ads in its Maps app. This sounds like Google Maps, and I don't like it. I never find these relevant, and they clutter up the user experience. Apple Maps now has a UI advantage (though not a data/search one, which matters more).

Apple is nickel-and-diming its customers. We spend thousands for their products and premium services like Apple One. We all know why: income must rise, and new firms are needed to scale. This will eventually backfire.

900K impressions, 8K clicks, and $$$ orders…

I recently started dropshipping. Now that I own my business and can charge it as a business expense, it feels less like money wasted if it doesn't work. I also made t-shirts to sell. I intended to open a t-shirt store and had many designs on a hard drive. I read that Tiktok advertising had a high conversion rate and low cost because they were new. According to many, the advertising' cost/efficiency ratio would plummet and become as bad as Google or Facebook Ads. Now felt like the moment to try Tiktok marketing and dropshipping. I work in marketing for a SaaS firm and have seen how poorly ads perform. I wanted to try it alone.

I set up $250 and ran advertising for a week. Before that, I made my own products, store, and marketing. In this post, I'll show you my process and results.

Setting up the store

Dropshipping is a sort of retail business in which the manufacturer ships the product directly to the client through an online platform maintained by a seller. The seller takes orders but has no stock. The manufacturer handles all orders. This no-stock concept increases profitability and flexibility.

In my situation, I used previous t-shirt designs to make my own product. I didn't want to handle order fulfillment logistics, so I looked for a way to print my designs on demand, ship them, and handle order tracking/returns automatically. So I found Printful.

I needed to connect my backend and supplier to a storefront so visitors could buy. 99% of dropshippers use Shopify, but I didn't want to master the difficult application. I wanted a one-day project. I'd previously worked with Big Cartel, so I chose them.

Big Cartel doesn't collect commissions on sales, simply a monthly flat price ($9.99 to $19.99 depending on your plan).

After opening a Big Cartel account, I uploaded 21 designs and product shots, then synced each product with Printful.

Developing the ads

I mocked up my designs on cool people photographs from placeit.net, a great tool for creating product visuals when you don't have a studio, camera gear, or models to wear your t-shirts.

I opened an account on the website and had advertising visuals within 2 hours.

Because my designs are simple (black design on white t-shirt), I chose happy, stylish people on plain-colored backdrops. After that, I had to develop an animated slideshow.

Because I'm a graphic designer, I chose to use Adobe Premiere to create animated Tiktok advertising.

Premiere is a fancy video editing application used for more than advertisements. Premiere is used to edit movies, not social media marketing. I wanted this experiment to be quick, so I got 3 social media ad templates from motionarray.com and threw my visuals in. All the transitions and animations were pre-made in the files, so it only took a few hours to compile. The result:

I downloaded 3 different soundtracks for the videos to determine which would convert best.

After that, I opened a Tiktok business account, uploaded my films, and inserted ad info. They went live within one hour.

The (poor) outcomes

As a European company, I couldn't deliver ads in the US. All of my advertisements' material (title, description, and call to action) was in English, hence they continued getting rejected in Europe for countries that didn't speak English. There are a lot of them:

I lost a lot of quality traffic, but I felt that if the images were engaging, people would check out the store and buy my t-shirts. I was wrong.

• 51,071 impressions on Day 1. 0 orders after 411 clicks

• 114,053 impressions on Day 2. 1.004 clicks and no orders

• Day 3: 987 clicks, 103,685 impressions, and 0 orders

• 101,437 impressions on Day 4. 0 orders after 963 clicks

• 115,053 impressions on Day 5. 1,050 clicks and no purchases

• 125,799 impressions on day 6. 1,184 clicks, no purchases

• 115,547 impressions on Day 7. 1,050 clicks and no purchases

• 121,456 impressions on day 8. 1,083 clicks, no purchases

• 47,586 impressions on Day 9. 419 Clicks. No orders

My overall conversion rate for video advertisements was 0.9%. TikTok's paid ad formats all result in strong engagement rates (ads average 3% to 12% CTR to site), therefore a 1 to 2% CTR should have been doable.

My one-week experiment yielded 8,151 ad clicks but no sales. Even if 0.1% of those clicks converted, I should have made 8 sales. Even companies with horrible web marketing would get one download or trial sign-up for every 8,151 clicks. I knew that because my advertising were in English, I had no impressions in the main EU markets (France, Spain, Italy, Germany), and that this impacted my conversion potential. I still couldn't believe my numbers.

I dug into the statistics and found that Tiktok's stats didn't match my store traffic data.

Looking more closely at the numbers

My ads were approved on April 26 but didn't appear until April 27. My store dashboard showed 440 visitors but 1,004 clicks on Tiktok. This happens often while tracking campaign results since different platforms handle comparable user activities (click, view) differently. In online marketing, residual data won't always match across tools.

My data gap was too large. Even if half of the 1,004 persons who clicked closed their browser or left before the store site loaded, I would have gained 502 visitors. The significant difference between Tiktok clicks and Big Cartel store visits made me suspicious. It happened all week:

• Day 1: 440 store visits and 1004 ad clicks

• Day 2: 482 store visits, 987 ad clicks

• 3rd day: 963 hits on ads, 452 store visits

• 443 store visits and 1,050 ad clicks on day 4.

• Day 5: 459 store visits and 1,184 ad clicks

• Day 6: 430 store visits and 1,050 ad clicks

• Day 7: 409 store visits and 1,031 ad clicks

• Day 8: 166 store visits and 418 ad clicks

The disparity wasn't related to residual data or data processing. The disparity between visits and clicks looked regular, but I couldn't explain it.

After the campaign concluded, I discovered all my creative assets (the videos) had a 0% CTR and a $0 expenditure in a separate dashboard. Whether it's a dashboard reporting issue or a budget allocation bug, online marketers shouldn't see this.

Tiktok can present any stats they want on their dashboard, just like any other platform that runs advertisements to promote content to its users. I can't verify that 895,687 individuals saw and clicked on my ad. I invested $200 for what appears to be around 900K impressions, which is an excellent ROI. No one bought a t-shirt, even an unattractive one, out of 900K people?

Would I do it again?

Nope. Whether I didn't make sales because Tiktok inflated the dashboard numbers or because I'm horrible at producing advertising and items that sell, I’ll stick to writing content and making videos. If setting up a business and ads in a few days was all it took to make money online, everyone would do it.

Video advertisements and dropshipping aren't dead. As long as the internet exists, people will click ads and buy stuff. Converting ads and selling stuff takes a lot of work, and I want to focus on other things.

I had always wanted to try dropshipping and I’m happy I did, I just won’t stick to it because that’s not something I’m interested in getting better at.

If I want to sell t-shirts again, I'll avoid Tiktok advertisements and find another route.

I'd spend hours composing articles only to find out they were useless. Twitter solved the problem.

Twitter is wrinkled, say critics.

Nope. Writing is different. It won't make sense until you write there.

Twitter is resurgent. People are reading again. 15-second TikToks overloaded our senses.

After nuking my 20,000-follower Twitter account and starting again, I wrote every day for 210 days.

I'll explain.

I came across the strange world of microblogging.

Traditional web writing is filler-heavy.

On Twitter, you must be brief. I played Wordle.

Twitter Threads are the most popular writing format. Like a blog post. It reminds me of the famous broetry posts on LinkedIn a few years ago.

Threads combine tweets into an article.

• Sharp, concise sentences

• No regard for grammar

• As important as the information is how the text looks.

Twitter Threads are like Michael Angelo's David monument. He chipped away at an enormous piece of marble until a man with a big willy appeared.

That's Twitter Threads.

I tried to remove unnecessary layers from several of my Wordpress blog posts. Then I realized something.

Tweeting from scratch is easier and more entertaining. It's quicker and makes you think more concisely.

Superpower: saying much with little words. My long-form writing has improved. My article sentences resemble tweets.

You never know what will happen.

Twitter's subcultures are odd. Best-performing tweets are strange.

Unusual trend: working alone and without telling anyone. It's a rebellion against Instagram influencers who share their every moment.

Early on, random thoughts worked:

My friend’s wife is Ukrainian. Her family are trapped in the warzone. He is devastated. And here I was complaining about my broken garage door. War puts everything in perspective. Today is a day to be grateful for peace.

Documenting what's happening triggers writing. It's not about viral tweets. Helping others matters.

There are numerous anonymous users.

Twitter uses pseudonyms.

You don't matter. On sites like LinkedIn, you must use your real name. Welcome to the Cyberpunk metaverse of Twitter :)

One daily piece of writing is a powerful habit.

Habits build creator careers. Read that again.

Twitter is an easy habit to pick up. If you can't tweet in one sentence, something's wrong. Easy-peasy-japanese.

Not what I tweeted, but my constancy, made the difference.

Daily writing is challenging, especially if your supervisor is on your back. Twitter encourages writing.

Tweets evolved as the foundation of all other material.

During my experiment, I enjoyed Twitter's speed.

Tweets get immediate responses, comments, and feedback. My popular tweets become newspaper headlines. I've also written essays from tweet discussions.

Sometimes the tweet and article were clear. Twitter sometimes helped me overcome writer's block.

I used to spend hours composing big things that had little real-world use.

Twitter helped me. No guessing. Data guides my coverage and validates concepts.

Test ideas on Twitter.

It took some time for my email list to grow.

Subscribers are a writer's lifeblood.

Without them, you're broke and homeless when Mark Zuckerberg tweaks the algorithms for ad dollars. Twitter has three ways to obtain email subscribers:

1. Add a link to your bio.

Twitter allows bio links (LinkedIn now does too). My eBook's landing page is linked. I collect emails there.

2. Start an online newsletter.

Twitter bought newsletter app Revue. They promote what they own.

I just established up a Revue email newsletter. I imported them weekly into my ConvertKit email list.

3. Create Twitter threads and include a link to your email list in the final tweet.

Write Twitter Threads and link the last tweet to your email list (example below).

Initial email subscribers were modest.

Numbers are growing. Twitter provides 25% of my new email subscribers. Some days, 50 people join.

Without them, my writing career is over. I'd be back at a 9-5 job begging for time off to spend with my newborn daughter. Nope.

Collect email addresses or die trying.

As insurance against unsubscribes and Zucks, use a second email list or Discord community.

What I still need to do

Twitter's fun. I'm wiser. I need to enable auto-replies and auto-DMs (direct messages).

This adds another way to attract subscribers. I schedule tweets with Tweet Hunter.

It’s best to go slow. People assume you're an internet marketer if you spam them with click requests.

A human internet marketer is preferable to a robot. My opinion.

210 days on Twitter taught me that. I plan to use the platform until I'm a grandfather unless Elon ruins it.

A completely operating application requires many processes and technical challenges. Implementing coding standards can make apps right, work, and faster.

With years of experience working in software houses. Many client apps are scarcely maintained.

Why are these programs "barely maintainable"? If we're used to coding concepts, we can probably tell if an app is awful or good from its codebase.

This is how I coded much of my app.

Make It Work

Before adopting any concept, make sure the apps are completely functional. Why have a fully maintained codebase if the app can't be used?

The user doesn't care if the app is created on a super server or uses the greatest coding practices. The user just cares if the program helps them.

After the application is working, we may implement coding principles.

You Aren’t Gonna Need It

As a junior software engineer, I kept unneeded code, components, comments, etc., thinking I'd need them later.

In reality, I never use that code for weeks or months.

First, we must remove useless code from our primary codebase. If you insist on keeping it because "you'll need it later," employ version control.

If we remove code from our codebase, we can quickly roll back or copy-paste the previous code without preserving it permanently.

The larger the codebase, the more maintenance required.

Keep It Simple Stupid

Indeed. Keep things simple.

Why complicate something if we can make it simpler?

Our code improvements should lessen the server load and be manageable by others.

If our code didn't pass those benchmarks, it's too convoluted and needs restructuring. Using an open-source code critic or code smell library, we can quickly rewrite the code.

Simpler codebases and processes utilize fewer server resources.

Don't Repeat Yourself

Have you ever needed an action or process before every action, such as ensuring the user is logged in before accessing user pages?

As you can see from the above code, I try to call is user login? in every controller action, and it should be optimized, because if we need to rename the method or change the logic, etc. We can improve this method's efficiency.

We can write a constructor/middleware/before action that calls is_user_login?

The code is more maintainable and readable after refactoring.

Each programming language or framework handles this issue differently, so be adaptable.

Clean Code

Clean code is a broad notion that you've probably heard of before.

When creating a function, method, module, or variable name, the first rule of clean code is to be precise and simple.

The name should express its value or logic as a whole, and follow code rules because every programming language is distinct.

If you want to learn more about this topic, I recommend reading https://www.amazon.com/Clean-Code-Handbook-Software-Craftsmanship/dp/0132350882.

Standing On The Shoulder of Giants

Use industry standards and mature technologies, not your own(s).

There are several resources that explain how to build boilerplate code with tools, how to code with best practices, etc.

I propose following current conventions, best practices, and standardization since we shouldn't innovate on top of them until it gives us a competitive edge.

Boy Scout Rule

What reduces programmers' productivity?

When we have to maintain or build a project with messy code, our productivity decreases.

Having to cope with sloppy code will slow us down (shame of us).

How to cope? Uncle Bob's book says, "Always leave the campground cleaner than you found it."

When developing new features or maintaining current ones, we must improve our codebase. We can fix minor issues too. Renaming variables, deleting whitespace, standardizing indentation, etc.

Make It Fast

After making our code more maintainable, efficient, and understandable, we can speed up our app.

Whether it's database indexing, architecture, caching, etc.

A smart craftsman understands that refactoring takes time and it's preferable to balance all the principles simultaneously. Don't YAGNI phase 1.

Using these ideas in each iteration/milestone, while giving the bottom items less time/care.

You can check one of my articles for further information. https://medium.com/life-at-mekari/why-does-my-website-run-very-slowly-and-how-do-i-optimize-it-for-free-b21f8a2f0162

https://medium.com/life-at-mekari/what-you-need-to-make-your-app-a-high-availability-system-tackling-the-technical-challenges-8896abec363f

Old power paradigm blocks new planetary paradigm

The difference between our reality and the media's reality is like a tale of two worlds. The greatest and worst of times, really.

Expanding information demands complex skills and understanding to separate important information from ignorance and crap. And that's just the start of determining the source's aim.

Trust who? We see people trust liars in public and then be destroyed by their decisions. Mistakes may be devastating.

Many give up and don't trust anyone. Reality is a choice, though. Same risks.

We must separate our needs and wants from reality. Needs and wants have rules. Greed and selfishness create an unlivable planet.

Culturally, we know this, but we ignore it as foolish. Selfish and greedy people obtain what they want, while others suffer.

We invade, plunder, rape, and burn. We establish civilizations by institutionalizing an exploitable underclass and denying its existence. These cultural lies promote greed and selfishness despite their destructiveness.

Controlling parts of society institutionalize these lies as fact. Many of each age are willing to gamble on greed because they were taught to see greed and selfishness as principles justified by prosperity.

Our cultural understanding recognizes the long-term benefits of collaboration and sharing. This older understanding generates an increasing tension between greedy people and those who see its planetary effects.

Survival requires distinguishing between global and regional realities. Simple, yet many can't do it. This is the first time human greed has had a global impact.

In the past, conflict stories focused on regional winners and losers. Losers lose, winners win, etc. Powerful people see potential decades of nuclear devastation as local, overblown, and not personally dangerous.

Mutually Assured Destruction (MAD) was a human choice that required people to acquiesce to irrational devastation. This prevented nuclear destruction. Most would refuse.

A dangerous “solution” relies on nuclear trigger-pullers not acting irrationally. Since then, we've collected case studies of sane people performing crazy things in experiments. We've been lucky, but the climate apocalypse could be different.

Climate disaster requires only continuing current behavior. These actions already cause global harm, but that's not a threat. These activities must be viewed differently.

Once grasped, denying planetary facts is hard to accept. Deniers can't think beyond regional power. Seeing planet-scale is unusual.

Decades of indoctrination defining any planetary perspective as un-American implies communal planetary assets are for plundering. The old paradigm limits any other view.

In the same way, the new paradigm sees the old regional power paradigm as a threat to planetary civilization and lifeforms. Insane!

While MAD relied on leaders not acting stupidly to trigger a nuclear holocaust, the delayed climatic holocaust needs correcting centuries of lunacy. We must stop allowing craziness in global leadership.

Nothing in our acknowledged past provides a paradigm for such. Only primitive people have failed to reach our level of sophistication.

Before European colonization, certain North American cultures built sophisticated regional nations but abandoned them owing to authoritarian cruelty and destruction. They were overrun by societies that saw no wrong in perpetual exploitation. David Graeber's The Dawn of Everything is an example of historical rediscovery, which is now crucial.

From the new paradigm's perspective, the old paradigm is irrational, yet it's too easy to see those in it as ignorant or malicious, if not both. These people are both, but the collapsing paradigm they promote is older or more ingrained than we think.

We can't shift that paradigm's view of a dead world. We must eliminate this mindset from our nations' leadership. No other way will preserve the earth.

Change is occurring. As always with tremendous transition, younger people are building the new paradigm.

The old paradigm's disintegration is insane. The ability to detect errors and abandon their sources is more important than age. This is gaining recognition.

The breakdown of the previous paradigm is not due to senile leadership, but to systemic problems that the current, conservative leadership cannot recognize.

Stop following the old paradigm.

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.