hoodtronik

Trellis 2 image-to-mesh + UltraTex PBR texturing (FLUX.2 Klein) in a simple GUI, running on a private ComfyUI backend. Windows + NVIDIA. Low VRAM (8GB) default, High VRAM mode for 16GB+.

Generate complete textured 3D assets (GLB) from sparse, unposed multi-view images. Joint 3D reconstruction + generation built on TRELLIS.2 and VGGT, with a Gradio web UI. Native Windows install using prebuilt CUDA wheels.

LAION's Chatterbox-based voice converter: give it a source recording and a short clip of a consenting target speaker, and it says the source content in the target voice while keeping the source performance and emotion. Gradio GUI, CUDA GPU, ~1.5 GB weights.

Local AI sound design from video (SFX + music bed, mixed and mastered)

Controllable video-to-audio Foley (Xiaomi, ACM MM 2026)

Tencent Hunyuan AuK — 1.5B speech generation & editing. Rewrite what is SAID in an existing recording (fix a mispronounced word) while keeping the original voice and delivery. Also zero-shot TTS, emotion/timbre/whisper edits, denoise, dereverb and speaker separation. MIT licensed.

Pixel-Aligned 3D Generation from Images (SIGGRAPH 2026). Generate high-fidelity 3D GLB assets with PBR textures from a single image, powered by the TRELLIS.2 backbone. Native Windows install using prebuilt CUDA wheels (Python 3.12 / torch 2.8 / CUDA 12.8).

Local LoRA trainer for MiniMax H3, Flux2 Klein 9B & Krea 2 — trains against the exact models you deploy. Desktop GUI + headless CLI. NVIDIA only (RTX 30/40/50, driver 555+).

Gradio UI for the official LTX-2 video model trainer — LoRA, audio-video, IC-LoRA, and full fine-tuning.

a state-of-the-art open-source model for fast feedforward 3D reconstruction from a single image, developed in collaboration between Tripo AI and Stability AI. https://huggingface.co/spaces/stabilityai/TripoSR

Pinokio launcher for LTX-Desktop-WanGP (local video generation with WanGP backend)

Train LoRA / LoHa / LoKr for Wan2.2, FLUX.2, Z-Image, HunyuanVideo, and more — one-click install of kohya-ss/musubi-tuner with its built-in Gradio GUI.

Convert a single 2D image into high-quality 3D Gaussian Splats (.ply / .splat), by TripoAI. NVIDIA GPU required.

Local web UI for orchestrating 3D generation pipelines via ComfyUI / Tripo / Tencent. https://github.com/visualbruno/3DGenStudio

Gradio UI for the Klippbok video dataset curation CLI — LoRA dataset prep for video models.

Kimodo: Kinematic Motion Diffusion Model. Generates high-quality 3D human and robot motions.

Reconstruct 3D Gaussian Splatting worlds from video using non-rigid alignment. Supports fast and extensive modes with 2DGS/3DGS rendering. https://github.com/lukasHoel/video_to_world

Video-to-Audio generation with Chain-of-Thought reasoning (ICLR 2026)

Zero-shot voice conversion, real-time voice conversion, and singing voice conversion. https://github.com/Plachtaa/seed-vc

Self-supervised Generative LM-based Voice Conversion