FLUX 1.0-dev is a cutting-edge, 12 billion parameter rectified flow transformer capable of generating highly detailed images from text descriptions, designed for flexible use in personal, scientific, and non-commercial applications.

While this model delivers exceptional results, it may fail to follow prompts precisely and can amplify biases. It’s unsuitable for factual outputs and restricted from certain harmful or illegal uses as outlined in its non-commercial license.

• vCPU - 4x
• RAM - 30 GB
• GPU - RTX 4090 (24GB)

DreamShaper 8 is a text-to-image generation model, fine-tuned from Stable Diffusion 1.5, designed to create artistic and detailed imagery across a wide range of styles, from realism to anime.

This model may struggle with highly photorealistic outputs compared to models specialized in that area, but it offers flexibility for both anime and realism with proper skills. Caution is advised for unintended outputs, including NSFW content.

• vCPU - 4x
• RAM - 12 GB
• GPU - RTX 3060 Ti (8GB)

Stable Diffusion XL (SDXL) is a latent diffusion model for text-to-image synthesis. The model is an advanced deep generative model designed to create high-quality 1024x1024 px images from textual descriptions.

The model was not trained to be factual or true representations of people or events, and therefore using the model to generate such content is out-of-scope for the abilities of this model.

• vCPU - 2x
• RAM - 30 GB
• GPU - RTX 4090 (24GB)

YOLOv8 is a state-of-the-art model from Ultralytics, designed for various tasks such as object detection, instance segmentation, image classification, and pose estimation.

This model excels in fast, accurate object detection and other computer vision tasks but has not been fully optimized for all use cases. Users should explore model training, validation, and customization to ensure it fits their specific application requirements.

• vCPU - 8x
• RAM - 16 GB
• GPU - RTX 4060 Ti (16GB)

Llama 3.1 is an instruction-tuned model by Meta with 8 billion parameters, optimized for multilingual dialogue and text generation.

Llama 3.1 is optimized for text-based tasks like chatbots, generating coherent, contextually relevant responses across multiple languages. However, users should exercise caution in unsupported languages or sensitive scenarios. Further fine-tuning for specific use cases is recommended to ensure ethical compliance and avoid misuse.

• vCPU - 8x
• RAM - 8 GB
• GPU - RTX 3060 (12GB)

Mistral 7B v0.1 is a pretrained generative language model with 7 billion parameters, optimized for text generation tasks.

As a base model, Mistral 7B v0.1 lacks built-in moderation or safety features, so users should proceed with caution when deploying it in sensitive contexts. Additional fine-tuning is strongly recommended for tasks that require stricter safety, compliance protocols, or specialized performance to ensure proper use.

• vCPU - 8x
• RAM - 16 GB
• GPU - RTX 3090 (24GB)

GPT2 is a transformers model that was pretrained on a large volume of raw English text with no human labelling. This model allows you to generate text based on a prompt, based on data from the training model. This is the smaller available model, with 124M parameters.

GPT2 can be used to expand a starter text prompt. As GPT2 has been trained on unfiltered content from the internet, it may contain biased and incorrect information. Content generated by GPT2 should be filtered and corrected by downstream studying.

• vCPU - 4x
• RAM - 8GB
• GPU - N/A

MPT-7B is a decoder style transformer trained on 1 Trillion tokens of English language text by MosaicML. This is the 7B-Storywriter variant of the model

This model is not intended for deployment without finetuning or guardrails for human interaction.

• vCPU - 4x
• RAM - 16GB
• GPU - RTX 3080 Ti (12GB)