rtx-50

Windows player that runs video, photos or GIFs through NVIDIA's DLSS 5 neural renderer, then compares the result with the original frame by frame. Local files or YouTube. Needs an RTX card and driver 610.47+. Community project, unsigned runtime. C++20, D3D12.

Windows prebuilt of llama.cpp combining Multi-Token Prediction (MTP) + TurboQuant KV cache compression + native sm_120 (Blackwell consumer GPU, FP4 tensor cores). For RTX 5060 Ti / 5070 / 5080 / 5090.

MiniMax H3 audio-video model on TensorFold NVFP4 kernels: a faster ComfyUI loader for one RTX 50-series GPU (2.3x stock at 20 steps, 81 s per 5 s 1344x768 video with Turbo + sparse attention on an RTX 5070 Ti)

Run llama.cpp with Multi-Token Prediction and TurboQuant on Windows using native sm_120 Blackwell support for RTX 50-series GPUs.

OBS Studio fork targeting periodic capture stutter, duplicate frames, and preview/recording skips with stable OBS FPS and no render or encoder overload; initially reproduced on RTX 50 series.