by calcuis
Open source · 4k downloads · 32 likes
Cosmos Predict2 GGUF is a quantized and optimized version of the Cosmos Predict2 AI model, available in two sizes (2 billion and 14 billion parameters). Designed to generate images from text, it seamlessly integrates with environments like ComfyUI and uses the GGUF format, which is lighter and compatible with modern tools. Its key strengths lie in its ability to produce detailed and coherent visuals from text descriptions while offering greater flexibility through compatibility with GGUF encoders and decoders. The model stands out for its ease of use, lack of heavy dependencies like safetensors files, and its adaptability to automated image generation workflows. Ideal for artists, developers, or content creators seeking to leverage generative AI without excessive technical complexity.
./ComfyUI/models/diffusion_models./ComfyUI/models/text_encoders./ComfyUI/models/vae

ex: extra state excluded; ex3: ex + pos_embedder x 3 excluded