Category
Explore by categories
Provides a DiT-based audio–video foundation model plus an official Python inference and LoRA trainer. Ships multiple production-ready pipelines (text/image/audio→video), checkpoints, and performance optimizations (FP8, distilled pipelines) for high-fidelity synchronized audio–video generation.
Generates controllable multilingual speech from text with nine predefined timbres and custom-voice control; supports voice design, quick voice cloning and low-latency streaming (first audio packet after a single character), suitable for real-time TTS and voice-design workflows.
Generates anime-style and other non-photorealistic illustrations from text prompts. A 2B-parameter diffusion base preview trained on millions of anime images (and ~800k non-anime art) and released under a non-commercial license; best used in ComfyUI around ~1MP resolution.
A 26M-parameter LLM distilled for reliable function-call generation on tiny devices, with open weights, local finetuning tooling, and a web playground for on-device testing. Pretrained at scale then post-trained on a single-shot function-call dataset for tool integration.
An instruction‑tuned Gemma 4 E4B multimodal model on Hugging Face that accepts text, images and audio and generates text; notable for 128K long context support, built-in thinking mode, and an on‑device‑friendly E4B architecture under an Apache‑2.0 license.
Identifies and surgically removes the internal activation directions that trigger refusal behavior in large language models, with one-click options on a HuggingFace Space or a local Python API. Combines multiple extraction methods (SVD, whitened SVD, sparse autoencoders), reversible steering, and analysis-informed verification to quantify capability and refusal trade-offs.
Community fine-tuned multimodal Qwen3.5-9B using Claude 4.6 distilled data to change the model's 'thinking' behavior; offers an uncensored 'heretic' flavor with image-text-to-text I/O, benchmark comparisons, and deployment notes for inference frameworks.
Generate text, images, video, audio and action/robot trajectories from combined text, image, video, audio and action inputs. A Mixture-of-Transformers omnimodal foundation model (Cosmos3‑Nano, 16B params) focused on Physical AI (robotics, AV, simulation) and optimized for NVIDIA GPU runtimes.
Instruction-tuned Gemma 4 31B multimodal model that generates text from text+image inputs with up to 256K context. Dense 31B variant optimized for vision-language understanding, long-context reasoning, and coding; Apache‑2.0 licensed.
Instruction-tuned Mixture-of-Experts multimodal model that generates text from text+image inputs while activating a 4B subset of parameters for faster inference; supports a 256K context window, multilingual vision-language tasks, and is available under Apache-2.0.
A distilled 26M-parameter encoder–decoder LLM for on-device function-calling and tool use. Uses a pure-attention Simple Attention Network, provides open weights and local finetuning, and targets high-throughput inference on the Cactus runtime.
Pretrained image-model checkpoint hosted on Hugging Face by Facebook (Meta) for vision experiments and transfer learning. Includes downloadable weights and metadata under CC BY‑NC 4.0 — suitable for research and prototyping but restricted for commercial use.