Browse

FLUX is a 12 billion parameter rectified flow transformer text-to-image model. The "Schnell" model is the fastest version of FLUX.1 which can generate high-quality images in fewer steps.

Version 10 of the popular JuggernautXL model, with improved prompt adherence from revamped labelling of training dataset.

SDXL fine tune with broad competence in art generation, animation, and photorealism, using Lightning to generate images very fast.

SDXL fine-tune optimizing for generating photorealistic people, animals, and landscapes.

Base 1024px text-to-image model using the SDXL architecture trained to produce results that are 3x more aesthetically pleasing compared to SDXL.

AnimateDiff motion model to generate videos using Stable Diffusion XL.

Official instruction-oriented fine-tune of Llama3 70B, recommended over the base model for most use cases.

Mixtral 8x7B fine-tuned on a large dataset of instructional text, yielding GPT 3.5 level performance in chat and generalist use cases.

Mixtral instruction-following fine-tune with strong coding abilities, trained on a dataset inspired by the Orca paper and intentionally unaligned / uncensored.

Llama 2 70B model fine-tuned to follow complex instructions using the WizardLM training methodology.

Llama 2 70B fine-tuned on over 300k curated GPT-4 outputs, known long high-quality outputs.

Instruct fine-tune of the Gemma 7B LLM from Google, developed using the same technology as their flagship Gemini models.

Widely used transformer model from OpenAI for SST (speech to text), automatic speech recognition (ASR), and speech translation.

Bark is a transformer text-to-audio model for creating multilingual speech and other audio such as music, background noise, and sound effects.

MusicGen is a text-to-music autoregressive transformer model capable of generating music samples conditioned on text descriptions or audio prompts.

Magnet is a single-stage, non-autoregressive transformer model for text-to-music and text-to-audio generation, with faster generation that MusicGen.

Seamless is a MMMMT (Massively Multilingual and Multimodal Machine Translation) model that can translate and generate audio in over 100 languages.

Vocos is a fast neural vocoder, Generative Adversarial Network (GAN) objective, designed to synthesize audio waveforms from acoustic features.