Apertus 70B: Truly Open - Swiss LLM by ETH, EPFL and CSCS

I built a Universal File Converter with Gradio & Hugging Face looking for feedback 🚀

Wan2.2-S2V-14B – audio-driven cinematic video generation model

grok-2 on Hugging Face

Qwen Image

LFM2 WebGPU

Qwen3-4B-Thinking-2507

Introducing Pivotal Token Search (PTS): Targeting Critical Decision Points in LLM Training

Unsupervised Model Improvement via Internal Coherence Maximization: Outperforming Human-Supervised Methods Through Self-Elicitation

Implemented the research paper “Memorizing Transformers” from scratch with my own additional modifications in architecture and customized training pipeline .