This is a 32B parameter reward model developed by NVIDIA based on the Qwen3 architecture. It is specifically used for reward scoring and principle alignment in reinforcement learning, helping to train AI systems that are safer and more in line with human values.

Natural Language ProcessingGguf

This model is based on GPT-OSS-20B and fine-tuned using the Unsloth reinforcement learning framework. The aim is to optimize inference efficiency and reduce vulnerabilities that occur during reinforcement learning from human feedback (RLHF) training. The fine-tuning process focuses on the robustness and efficiency of alignment, ensuring that the model maintains inference depth without incurring excessive computational overhead.

Natural Language ProcessingTransformersEnglish

The Skywork Critic series of models are developed by the Skywork AI Alignment Team. It includes two advanced critic models, 70B and 8B, which are good at paired preference evaluation and can conduct detailed comparisons of input text pairs to judge their relative quality or applicability.

Natural Language ProcessingPytorch

Mini Magnum is a streamlined version of magnum-72b-v1, aiming to replicate the text quality of the Claude 3 series (especially Sonnet and Opus). It is fine-tuned based on Mistral-Nemo-Base-2407, and a new general instruction dataset is added to improve coherence and alignment.

Natural Language ProcessingTransformersMultiple Languages

GEITje 7B Ultra is a Dutch conversation model based on Mistral, optimized for Dutch dialogue through DPO and AI feedback alignment.

Natural Language ProcessingTransformersOther

Taiyi-Diffusion-XL is a bilingual diffusion model that focuses on enhancing the Chinese text-to-image generation ability while retaining English comprehension ability. Based on the Stable-Diffusion-XL architecture and trained with a high-quality bilingual dataset, it performs excellently in terms of image quality, diversity, and text alignment, providing new options for AIGC and digital art creation.

MultimodalDiffusersOther