This is the third version of the monolingual encoder-decoder language model released by the HPLT project, specifically designed for German. It is a text-to-text transformer model trained based on the denoising objective, following the NorT5 architecture settings, and using a tokenizer trained with German HPLT data.

Natural Language ProcessingPytorchMultiple Languages

SRPO is a human preference alignment method for diffusion models. Through Direct-Align technology and semantic relative preference optimization, it significantly improves the realism and aesthetic quality of the FLUX.1-dev model, and solves the problems of high computational cost in multi-step denoising and dependence on offline reward fine-tuning.

Computer VisionDiffusers

FRIDA is a general-purpose text embedding model fine-tuned with full parameters based on the T5 denoising architecture, supporting Russian and English text processing.

Natural Language ProcessingSentence-transformersMultiple Languages

The Qwentile Λ 2.5 32B Instruct Model is the result of standardized denoising Fourier interpolation fusion of multiple excellent models, combining thinking capabilities and creative output.

Natural Language ProcessingTransformersEnglish

A standardized denoising Fourier interpolation hybrid model based on the Qwen series, renowned for high-quality responses and emotional resonance as a thinking-oriented large language model

Natural Language ProcessingTransformersEnglish

An audio denoising and voice enhancement model based on Pytorch, which effectively removes audio noise and improves voice clarity

Audio ProcessingSafetensors

A Pokemon image generation model based on PyTorch and Diffusers library, using Denoising Diffusion Probabilistic Model (DDPM) architecture

Computer VisionDiffusers

Badger is a Llama3 8B instruction model generated through recursive maximum pairwise disjoint normalized denoising Fourier interpolation method, incorporating features from multiple high-quality models.

Natural Language ProcessingTransformers

This is an unconditional image generation model based on DDPM (Denoising Diffusion Probabilistic Model), specifically fine-tuned for butterfly images.

Computer VisionDiffusers

This is a speech enhancement model based on the SepFormer architecture, specifically designed for denoising tasks. It was trained on the Microsoft DNS-4 dataset and supports audio processing at a 16kHz sampling rate.

Audio ProcessingPytorchMultiple Languages

The SD-XL 1.0 Refiner Model is an image generation model developed by Stability AI, designed to enhance the quality of images generated by the SDXL base model, with particular expertise in the final denoising step.

Computer VisionDiffusers

This is a speech enhancement model based on the SepFormer architecture, specifically designed for denoising German speech in rescue scenarios, performing excellently at a 16kHz sampling rate.

Audio ProcessingPytorchGerman

This is a diffusion model for generating whole slide images (WSI), implemented based on the Denoising Diffusion Probabilistic Model (DDPM) architecture.

Computer VisionTensorboardMultiple Languages

Russian pre-trained language model based on T5 architecture, employing a UL2-like mixed training strategy with 7 denoising tasks, 1.7 billion parameters

Natural Language ProcessingTransformersOther

DDPM is a high-quality image generation model based on diffusion probabilistic models, generating images through a progressive denoising process.

Computer VisionDiffusers

A high-quality image generation model based on diffusion probabilistic models, generating images through a progressive denoising process.

Computer VisionDiffusers

DDPM is a high-quality image generation model based on diffusion probabilistic models, inspired by non-equilibrium thermodynamics, which generates images through a progressive denoising process.

Computer VisionDiffusers

A simplified diffusion model based on U-Net architecture for teaching demonstrations on image denoising and generation

Computer VisionTensorflow

A closed-book QA model based on the T5 architecture, pre-trained with denoising and salient span masking objectives, suitable for QA tasks without external knowledge sources.

Natural Language ProcessingTransformersEnglish

plT5 is a language model based on the T5 architecture, specifically trained on Polish corpus and optimized for the original T5 denoising tasks.

Natural Language ProcessingTransformersOther