Stable Diffusion NSFW: Create Jaw-Dropping AI Art NOW!

The creation of AI-generated imagery, specifically stable diffusion nsfw content, has rapidly expanded with tools like AUTOMATIC1111’s web UI. This method allows users to generate images based on text prompts. The ethical considerations surrounding the use of Stable Diffusion, a latent text-to-image diffusion model developed by Stability AI, is significant. Furthermore, the implementation of content filters designed to prevent the creation of certain image types is often bypassed to achieve the desired stable diffusion nsfw results. OpenAI’s image generation capabilities offer a contrast to Stable Diffusion’s approach regarding NSFW limitations, and this difference has significant implications for artists and developers alike who seek to explore stable diffusion nsfw.

AI Art’s Bold New Frontier: NSFW and Stable Diffusion

The realm of digital art is undergoing a seismic shift, fueled by the rapid advancement of artificial intelligence. AI art, once a niche curiosity, has exploded into mainstream consciousness, captivating artists, technologists, and the general public alike. This burgeoning field presents both unprecedented opportunities and complex ethical challenges.

One particularly intriguing and often controversial area within AI art is the generation of Not Safe For Work (NSFW) content, especially through powerful tools like Stable Diffusion. This exploration into the intersection of AI and adult-oriented imagery has ignited debates about artistic expression, responsible technology use, and the very definition of art itself.

Defining AI Art: A New Creative Medium

At its core, AI art refers to artworks created, generated, or significantly enhanced through the use of artificial intelligence algorithms. These algorithms, often based on machine learning, can analyze vast datasets of images, styles, and techniques, allowing them to generate entirely new pieces or modify existing ones in remarkable ways.

The increasing popularity of AI art stems from several factors: its accessibility, its potential for innovation, and its ability to democratize the creative process. Tools like Stable Diffusion empower individuals, regardless of their artistic background, to bring their visions to life through the power of AI.

The Allure and Controversy of NSFW AI Art

The application of Stable Diffusion to create NSFW content has generated significant interest, driven by the same factors that fuel the broader AI art movement: accessibility, novelty, and creative freedom. Users can now explore their fantasies, experiment with different artistic styles, and generate highly personalized imagery with relative ease.

However, this newfound power comes with a host of ethical and societal concerns. The creation of NSFW AI art raises questions about:

• Consent: Can AI-generated images depict realistic likenesses of real people without their explicit consent?
• Exploitation: Could AI be used to create non-consensual or exploitative content?
• Misinformation: How can we prevent the use of AI to generate deepfakes or other forms of harmful misinformation?
• Representation: How do we ensure fair and responsible representation within AI-generated NSFW content, avoiding harmful stereotypes?

These are just some of the critical questions that must be addressed as the field continues to evolve. The potential for misuse is undeniable, and responsible development and usage are paramount.

A Guide to Responsible Creation

This exploration serves as a guide for navigating the complex landscape of NSFW AI art using Stable Diffusion. Our primary goal is to empower users with the knowledge and tools necessary to create compelling and imaginative content while prioritizing safety, ethics, and responsible creation.

We will delve into prompt engineering, fine-tuning techniques, and the intricacies of safety filters and community guidelines. However, and most importantly, we will consistently emphasize the ethical considerations that must inform every stage of the creative process. Creating NSFW AI art is a powerful capability, and with great power comes great responsibility. We are committed to fostering a community of creators who are both skilled and ethically mindful.

The allure of generating personalized digital creations naturally leads us to examine the powerhouse behind much of today’s AI art: Stable Diffusion. Understanding its architecture, even at a high level, empowers users to wield it more effectively and appreciate the intricate processes unfolding beneath the surface.

Stable Diffusion Deconstructed: The Engine That Drives AI Art

Stable Diffusion has rapidly become a cornerstone of the AI art revolution, largely due to its ability to translate textual descriptions into stunningly detailed images. This section aims to demystify Stable Diffusion, explaining its core functionalities in an accessible manner, focusing on how it empowers users to create AI art.

Stable Diffusion: A Text-to-Image Maestro

At its heart, Stable Diffusion is a text-to-image model.

This means it takes a text prompt as input and generates a corresponding image as output. Unlike earlier AI models that operated directly on pixel data, Stable Diffusion leverages a more efficient process known as latent diffusion. This allows it to generate high-resolution images with remarkable speed and detail.

The Magic Behind the Image: Latent Diffusion Explained

The term "latent diffusion" might sound complex, but the underlying concept is relatively straightforward.

Instead of directly manipulating pixels, Stable Diffusion operates in a latent space. Think of this as a compressed representation of the image, capturing the essential features and characteristics.

The process involves several key steps:

1. Encoding: The input text prompt is encoded into a numerical representation that the model can understand.

2. Diffusion: The model starts with random noise in the latent space. Over multiple steps, it gradually removes this noise, guided by the encoded text prompt.

3. Decoding: Finally, the refined latent representation is decoded back into a pixel-based image, resulting in the final artwork.

By operating in the latent space, Stable Diffusion significantly reduces the computational resources needed for image generation, making it accessible to a wider range of users.

This approach allows for generating complex, high-resolution images faster than previous methods.

Automatic1111: Your Gateway to Stable Diffusion

While Stable Diffusion provides the engine, tools like Automatic1111 (often referred to as the Stable Diffusion Web UI) provide the user-friendly interface.

Automatic1111 simplifies the entire process of generating AI art, making it accessible even to those without extensive technical expertise.

Simplifying the User Experience

Automatic1111 provides a graphical user interface (GUI) that streamlines the workflow.

Users can easily input text prompts, adjust parameters, and generate images with just a few clicks. It offers a range of features, including:

• Prompt Management: Easy input and editing of text prompts.

• Parameter Tuning: Adjust settings like sampling methods, number of steps, and image resolution.

• Model Selection: Choose from a variety of pre-trained models and custom fine-tuned models.

• Real-time Preview: See the generated image as it evolves.

Accessibility and Customization

Automatic1111’s open-source nature and active community have led to a wealth of extensions and customizations.

Users can install plugins to add new features, integrate with other tools, and further personalize their AI art creation experience.

This accessibility and customization have made Automatic1111 the de facto standard for many Stable Diffusion users.

Its intuitive interface and extensive features make it an indispensable tool for anyone looking to explore the world of AI-generated art.

The allure of generating personalized digital creations naturally leads us to examine the powerhouse behind much of today’s AI art: Stable Diffusion. Understanding its architecture, even at a high level, empowers users to wield it more effectively and appreciate the intricate processes unfolding beneath the surface.

Navigating the NSFW Landscape: Boundaries, Ethics, and Responsibility

The rise of AI art, particularly in the realm of Not Safe For Work (NSFW) content, presents a complex web of ethical and legal considerations. It is crucial to navigate this landscape with a strong sense of responsibility, understanding the potential implications of generating and consuming such content.

Defining NSFW in the Context of AI Art

NSFW content, in the context of AI-generated art, encompasses a wide range of depictions. These depictions include explicit sexual acts, nudity, and other potentially offensive or disturbing material.

It is important to note that the definition of NSFW can be subjective and culturally dependent. Therefore, it is vital to be mindful of different perspectives and sensitivities when creating and sharing AI art.

Ethical Considerations and Potential Risks

The creation of NSFW AI art raises several critical ethical considerations. One of the most pressing is the potential for misuse, particularly in the creation of deepfakes or non-consensual imagery.

Deepfakes involve the manipulation of existing images or videos to depict individuals in compromising or fabricated situations. The creation and distribution of such content without consent can have devastating consequences for the individuals involved.

Another ethical concern is the perpetuation of harmful stereotypes. AI models are trained on vast datasets, which may contain biases that are reflected in the generated output. This can lead to the creation of images that reinforce negative stereotypes related to gender, race, or other protected characteristics.

The lack of clear legal frameworks surrounding AI-generated content further complicates the ethical landscape. It is essential to be aware of the potential legal implications of creating and sharing NSFW AI art, particularly in relation to copyright, defamation, and privacy laws.

The Importance of Responsible Creation and Consumption

Given the ethical and legal complexities surrounding NSFW AI art, it is imperative to approach its creation and consumption with a strong sense of responsibility. This involves being mindful of the potential impact of your actions on others and taking steps to mitigate potential harm.

Here are some key principles for responsible creation and consumption:

• Obtain Consent: Never create or share AI-generated content that depicts real individuals without their explicit consent.
• Avoid Creating Harmful Content: Refrain from creating content that promotes violence, discrimination, or exploitation.
• Be Mindful of Stereotypes: Be aware of the potential for AI models to perpetuate harmful stereotypes and take steps to mitigate this risk.
• Respect Copyright and Intellectual Property: Ensure that you have the necessary rights to use any copyrighted material in your AI-generated art.
• Consume Responsibly: Be critical of the content you consume and avoid supporting the creation or distribution of harmful or unethical AI art.
• Consider Watermarking: Implement watermarking to identify images as AI-generated to prevent potential misuse or misrepresentation.

Mitigating the Risk of Non-Consensual Imagery

One of the most significant concerns in the NSFW AI art space is the potential creation of non-consensual content. To mitigate this risk, creators should take the following steps:

• Avoid Using Real Names or Identifiable Features: When generating images, avoid using real names or other identifying information that could be used to target specific individuals.
• Implement Robust Filtering Mechanisms: Utilize filtering mechanisms to prevent the generation of content that depicts minors or involves non-consensual acts.
• Be Prepared to Respond to Abuse: Have a clear plan for responding to reports of abuse or misuse of your AI art platform.

Addressing Bias in AI Models

AI models are trained on data, and if that data reflects societal biases, the AI will as well. To combat this:

• Curate Training Data Carefully: When fine-tuning models, ensure the datasets used are diverse and representative to minimize the amplification of biases.
• Actively Monitor Output: Regularly review generated content for biased outputs and adjust prompts or fine-tuning parameters accordingly.
• Support Research into Bias Mitigation: Encourage and support research efforts aimed at identifying and mitigating bias in AI models.

Navigating the NSFW landscape in AI art requires a careful balance of creativity, ethical awareness, and legal understanding. By embracing responsible creation and consumption practices, we can harness the power of AI art while mitigating its potential risks and ensuring a safer, more ethical digital environment.

The previous sections highlighted the ethical considerations inherent in generating NSFW AI art. Acknowledging these boundaries allows us to explore the creative process itself. The power to generate explicit imagery hinges on the ability to communicate effectively with the AI – and that communication happens through prompts.

Mastering the Prompt: Crafting Compelling NSFW AI Art

Prompt engineering is the linchpin of AI art creation. It’s the art and science of crafting text prompts that guide AI models like Stable Diffusion to generate specific images.

For NSFW AI art, effective prompting is even more crucial. It dictates the difference between a blurred mess and a compelling, visually striking image that aligns with your vision.

The Importance of Prompt Engineering in NSFW AI Art

In essence, a prompt is a set of instructions for the AI. It describes the desired content, style, and composition of the image.

A well-crafted prompt acts as a precise blueprint. It allows you to shape the AI’s output and bring your creative ideas to life.

Especially when creating NSFW content, clarity is key. Ambiguous or vague prompts will lead to unpredictable results.

This often requires many iterations to hone in on the desired characteristics.

Therefore, understanding and mastering prompt engineering is essential for anyone seeking to create high-quality NSFW AI art.

Tips and Tricks for Effective NSFW Prompts

Crafting effective prompts involves a combination of detail, specificity, and artistic language. Here are several tips and tricks to help you elicit the desired results from Stable Diffusion:

• Be Specific: Avoid vague terms. Instead of “a woman,” try “a slender woman with long flowing red hair, wearing a revealing dress.”
• Use Descriptive Adjectives: Add detail by using descriptive adjectives that clarify the scene, subject, and mood. Think "sultry," "glowing," "intricate," "detailed," and "realistic."

• **Specify the Style: Define the artistic style. Options include "photorealistic," "anime," "fantasy," "cyberpunk," and "art nouveau." This helps to steer the AI toward a specific aesthetic.

• **Consider the Lighting: Lighting dramatically affects the mood and appearance of an image. Specify lighting conditions such as “soft lighting,” “backlit,” “dramatic shadows,” or “golden hour.”

• **Add Negative Prompts: Use negative prompts to exclude unwanted elements from the image. For example, if you don’t want distorted limbs, add "deformed, distorted limbs" to the negative prompt.

• **Use Keywords for Composition: Guide the AI on how to compose the shot. Use terms like "close-up," "full body shot," "wide angle," or "birds-eye view."

• **Experiment with Different Models and Seeds: Different Stable Diffusion models (e.g., Realistic Vision, Deliberate) will interpret prompts differently. Experiment with different seeds to explore variations of the same prompt.

• **Embrace Iteration: Don’t be afraid to tweak and refine your prompts. Iterate on your initial results, adding more detail or adjusting parameters until you achieve the desired outcome.

Examples of Effective NSFW Prompts

Here are examples of prompts designed to generate specific NSFW images, with an explanation of why they work:

Example 1: Realistic Portrait

• Prompt: “Photorealistic portrait of a young woman with piercing blue eyes, long flowing blonde hair, wearing a sheer lingerie set, soft lighting, bokeh background, 8k, highly detailed, intricate details”
• Explanation: This prompt specifies realism, age, eye color, hair style, clothing, lighting, and level of detail. These details combine to instruct the AI to create a high-quality portrait with a defined aesthetic.

Example 2: Fantasy Illustration

• Prompt: “A seductive succubus with crimson skin and large bat wings, perched atop a gothic tower overlooking a dark city, moonlight, dramatic shadows, fantasy art, intricate details, glowing eyes”
• Explanation: This prompt leverages descriptive language ("seductive succubus," "crimson skin," "gothic tower") and stylistic keywords ("fantasy art," "dramatic shadows," "glowing eyes") to generate a striking and fantastical image.

Example 3: Cyberpunk Scene

• Prompt: “A cyborg woman with neon pink hair and cybernetic implants, standing in a rain-slicked alleyway, wearing a revealing latex outfit, cyberpunk, dystopian, neon lights, highly detailed, 8k”
• Explanation: This prompt uses cyberpunk-specific terminology ("cyborg," "cybernetic implants," "latex outfit," "neon lights") to create a gritty, futuristic aesthetic. The added details and resolution instructions further enhance the image quality.

Important Considerations:

• Ethical Boundaries: Remember to avoid prompts that depict non-consensual acts, child exploitation, or any form of harmful or illegal content.
• Platform Restrictions: Be aware of the content policies of any platform where you intend to share your AI-generated art. Respect these policies and avoid generating content that violates their terms of service.
• Responsible Creation: Always create and consume NSFW AI art responsibly, with a focus on consent, respect, and avoiding harm.

Mastering prompt engineering is an ongoing process. But with practice, experimentation, and a strong understanding of ethical boundaries, you can unlock the full potential of Stable Diffusion and create compelling, visually stunning NSFW AI art.

The power to generate compelling NSFW art through clever prompting is undeniable.

However, the journey doesn’t end with the initial image generation. For those seeking to truly realize their artistic vision, the realm of fine-tuning awaits.

Fine-Tuning for Perfection: Achieving Your Artistic Vision

AI image generation, especially in the nuanced world of NSFW art, is often an iterative process. The initial output from Stable Diffusion, while potentially impressive, might not perfectly align with your creative aspirations.

This is where fine-tuning techniques come into play, offering a path to refine and enhance AI-generated images, allowing you to achieve your artistic vision with greater precision.

Understanding Model Fine-Tuning

At its core, model fine-tuning involves taking a pre-trained AI model, like Stable Diffusion, and training it further on a specific dataset. This dataset should reflect the desired style, subject matter, or characteristics you want the model to emphasize.

Think of it as teaching the AI to "specialize" in a particular area. This is a more involved process than simple prompting, but it yields more direct control.

The purpose of fine-tuning is to imbue the model with a deeper understanding of your artistic preferences.

By exposing it to a curated collection of images, you can guide it to produce outputs that are more consistent with your personal style and desired aesthetic.

LoRA: A More Accessible Approach

While full model fine-tuning can be resource-intensive and technically challenging, LoRA (Low-Rank Adaptation) provides a more accessible alternative.

LoRA allows you to train smaller, more manageable "adapter" layers that are added to the existing Stable Diffusion model.

These adapter layers learn to modify the model’s behavior in subtle but significant ways, allowing you to inject specific styles or characteristics without retraining the entire model from scratch.

This makes LoRA a practical choice for users with limited computational resources or those who are new to fine-tuning techniques.

LoRA models are also easier to share and distribute, fostering a vibrant community of artists and creators who can readily experiment with different styles and aesthetics.

ControlNet: Precise Control Over Image Generation

ControlNet represents another powerful tool in the fine-tuning arsenal. It offers unprecedented control over the composition, details, and artistic style of AI-generated images.

Unlike simple prompting, which relies on textual descriptions, ControlNet uses input images or other control signals to guide the image generation process.

For example, you can use a sketch, a segmentation map, or even a depth map to dictate the overall structure and layout of the generated image.

ControlNet and NSFW Art: Examples

Imagine wanting to create an NSFW image featuring a specific pose or anatomical detail. With ControlNet, you can provide a reference image with the desired pose, and the AI will generate a new image that adheres to that pose while incorporating other elements from your prompt.

• Pose Control: Use a simple stick figure drawing to define the pose of the subject.

• Canny Edge Detection: Guide the AI to follow specific lines and contours from an existing image.

• Depth Maps: Recreate 3D structures from existing images.

The ability to exert such precise control opens up a world of possibilities for NSFW art creation, allowing you to realize even the most complex and nuanced artistic visions.

By combining creative prompting with the targeted control offered by ControlNet, you can achieve a level of detail and precision that was previously unattainable.

Safety Filters and Community Guidelines: Navigating the Rules of Engagement

The creative potential of AI art is vast, but it exists within an ecosystem of rules and restrictions. Understanding and navigating safety filters and community guidelines is critical for responsible AI art creation, especially within the NSFW realm.

Understanding Safety Filters

AI art platforms, including those that utilize Stable Diffusion, often employ safety filters designed to prevent the generation of harmful or inappropriate content. These filters act as gatekeepers, attempting to identify and block prompts or outputs that violate pre-defined policies.

These policies can include restrictions on:

• Explicit sexual content involving minors
• Hate speech or discriminatory imagery
• Depictions of violence or illegal activities
• Personally identifiable information (PII)

Safety filters typically work by analyzing both the text prompt and the generated image. They utilize machine learning models trained to recognize patterns and features associated with prohibited content. When a potential violation is detected, the generation process may be halted, or the output may be censored.

The Nuances of Safety Filters

It’s important to recognize that safety filters are not perfect.

They can be overly sensitive, flagging harmless content due to ambiguous wording or unintentional associations. Conversely, they may be bypassed by cleverly crafted prompts that circumvent the filter’s detection mechanisms.

The effectiveness of safety filters varies across different platforms. Some platforms implement more stringent filters than others, reflecting their specific content policies and risk tolerance.

The Importance of Adhering to Community Guidelines

Beyond the automated safety filters, most AI art platforms also have community guidelines that outline acceptable behavior and content standards. These guidelines provide a broader framework for responsible use and aim to foster a positive and respectful community environment.

These guidelines typically cover:

• Prohibitions on harassment, bullying, and hate speech
• Restrictions on sharing illegal or harmful content
• Requirements to respect intellectual property rights
• Rules against spamming or manipulating the platform

Violating community guidelines can result in a range of consequences, from warnings and content removal to account suspension or permanent banishment.

The Spirit of the Guidelines

Adhering to community guidelines is not simply about avoiding penalties. It’s about contributing to a healthy and sustainable creative ecosystem.

By respecting the rules and engaging responsibly, users help to ensure that these platforms remain safe, inclusive, and productive spaces for artistic expression.

Strategies for Responsible Navigation

While the desire to push creative boundaries is natural, it’s essential to do so within the bounds of the law, ethics, and platform policies.

Here are some strategies for navigating safety filters and community guidelines responsibly:

• Familiarize yourself with the specific rules of each platform. Don’t assume that the same standards apply everywhere.
• Exercise caution when crafting prompts. Avoid language that could be interpreted as promoting violence, hate speech, or illegal activities.
• Be mindful of the potential impact of your creations. Consider whether your work could be offensive, harmful, or exploitative to others.
• Respect intellectual property rights. Avoid generating images that infringe on existing copyrights or trademarks.
• Prioritize ethical considerations over technical loopholes. Just because you can bypass a safety filter doesn’t mean you should.

Circumventing Censorship vs. Respecting Boundaries

The desire to circumvent safety filters can arise from a genuine interest in exploring artistic themes that are often censored. However, it’s crucial to distinguish between legitimate artistic expression and the creation of harmful or illegal content.

Bypassing filters to create content that exploits, abuses, or endangers others is never justifiable. Responsible AI artists should focus on pushing creative boundaries within ethical and legal constraints.

Respecting Artist’s Rights

Generating content inspired by another artist’s work is different from generating near-identical copies. It’s important to respect copyright law and avoid directly replicating existing artworks without permission. Credit the artist whenever inspiration is drawn from an artist’s style or content.

By following these guidelines, users can harness the power of AI art while upholding ethical standards and contributing to a safe and respectful creative community.

Ethical Considerations and Responsible AI Art Creation: A Deeper Dive

The creation of NSFW AI art brings forth a complex web of ethical considerations that demand careful scrutiny. Beyond technical skills and creative vision, lies a responsibility to navigate the potential pitfalls associated with generating and sharing such content. This section provides a more in-depth exploration of these ethical dimensions.

The Ethical Landscape of NSFW AI Art

NSFW AI art, by its very nature, often pushes the boundaries of societal norms and expectations. Therefore, understanding the ethical implications is paramount.

Consent and Representation

Consent is arguably the most critical ethical consideration. AI can create images that resemble real people.

Therefore, it’s vital to ensure that the creation of such images does not violate anyone’s right to privacy or misrepresent them without their explicit permission. This becomes even more crucial when the generated content is sexually suggestive or explicit.

Representation is another key aspect. AI models are trained on vast datasets, and biases present in these datasets can be amplified in the generated images.

This can lead to the perpetuation of harmful stereotypes related to gender, race, and sexuality. Creators must be mindful of these potential biases and strive to create diverse and inclusive representations.

The Specter of Deepfakes

Deepfakes, AI-generated media that convincingly impersonates real individuals, present a significant ethical challenge. The creation of non-consensual deepfake pornography can inflict profound emotional and reputational harm on the targeted individuals.

The technology allows for the fabrication of explicit content featuring someone without their knowledge or agreement. This can have devastating consequences.

It is imperative to avoid using AI art tools to create deepfakes without explicit consent, and to understand the potential legal repercussions of doing so.

Potential for Misuse and Risk Mitigation

The power of AI image generation also comes with the potential for misuse. Safeguarding against this requires vigilance and a proactive approach.

Non-Consensual Content and Harassment

AI art tools can be used to generate non-consensual intimate images or create content that contributes to online harassment. This type of misuse can have severe emotional and psychological consequences for the victims.

Creators must be aware of this potential and actively avoid using AI to create content that could be used to harass, threaten, or defame individuals.

Amplifying Harmful Stereotypes

As mentioned earlier, biases present in the training data can result in the perpetuation of harmful stereotypes. This is especially relevant in the context of NSFW content, where stereotypes related to gender, race, and sexuality are often prevalent.

Creators should strive to create diverse and inclusive representations, actively challenging existing stereotypes and promoting positive portrayals.

Mitigating Risks

Mitigating these risks requires a multi-pronged approach. It includes:

• Self-awareness: Recognizing one’s own biases and actively working to counter them.
• Critical evaluation: Scrutinizing the generated images for potential biases and harmful representations.
• Responsible prompting: Crafting prompts that promote diversity and inclusivity, rather than perpetuating stereotypes.
• Platform vigilance: Reporting instances of misuse or harmful content on AI art platforms.

Promoting Responsible Creation and Sharing Practices

The ethical use of AI for NSFW art extends beyond the creation process. It also encompasses how the content is shared and consumed.

Obtaining Consent

Whenever possible, obtaining explicit consent is paramount, especially when the generated images depict real individuals. This may not always be feasible in the context of AI art, but striving for consent should be a guiding principle.

Avoiding Harmful Content

Creators should actively avoid creating content that is likely to cause harm. This includes content that promotes violence, hate speech, or the exploitation of vulnerable individuals.

Consider the impact of the content on others. Ensure it does not contribute to a climate of fear or discrimination.

Responsible Sharing

When sharing NSFW AI art, it’s crucial to be mindful of the audience and the platform’s guidelines. Use appropriate content warnings and avoid sharing content in contexts where it is likely to be unwanted or offensive.

Consider the potential impact of the content on viewers and take steps to minimize the risk of harm.

By embracing these ethical considerations and promoting responsible creation and sharing practices, we can harness the creative potential of AI art while mitigating its potential harms.

Alright, go wild and create some awesome (and maybe a little naughty) art! Just remember to be mindful of the ethical stuff while you’re exploring the world of stable diffusion nsfw!