Anima-base-v1.0-GGUF

This repository contains GGUF weights for Anima-base-v1.0, a 2B parameter diffusion model based on the NVIDIA Cosmos architecture.

These weights have been quantized using a specialized pipeline to ensure maximum compatibility and image quality on consumer hardware.

🚀 Model Details

• Architecture: Cosmos (Diffusion)
• Parameters: 2 Billion
• Original Model: circlestone-labs/Anima
• Quantization: GGUF (K-Quants)

🛠️ Quantization Process

To ensure the highest possible quality and avoid the "BF16 noise bug" common in diffusion quantization, the following steps were taken:

1. FP32 Casting: The original BF16 safetensors were cast to FP32 before conversion.
2. Architecture Patching: Used a patched version of llama.cpp (based on build b3962) with the city96 diffusion patch to support the cosmos architecture.
3. Precision Preservation: Sensitive layers (normalizations and biases) were kept in F32 to prevent color banding and generation artifacts.

💾 Available Quants

File Name	Quant Type	Size	Description
anima-base-v1.0-Q8_0.gguf	Q8_0	~2.1 GB	Near-perfect fidelity, recommended for 16GB+ RAM.
anima-base-v1.0-Q6_K.gguf	Q6_K	~1.6 GB	Excellent balance of speed and quality.
anima-base-v1.0-Q5_K_M.gguf	Q5_K_M	~1.4 GB	The "Sweet Spot" for most users.
anima-base-v1.0-Q4_K_M.gguf	Q4_K_M	~1.2 GB	Highly efficient, great for low-VRAM/RAM.
anima-base-v1.0-Q3_K_M.gguf	Q3_K_M	~1.0 GB	Maximum compression.

🧩 Required Components

To use these GGUF weights, you still need the following parts from the original Anima release:

1. Text Encoder: qwen_3_06b_base.safetensors
2. VAE: qwen_image_vae.safetensors

🎮 How to Use

ComfyUI

1. Install ComfyUI-GGUF custom nodes.
2. Place the GGUF files in ComfyUI/models/unet/.
3. Use the Unet-Loader (GGUF) node to load the model.
4. Ensure you are using the correct Cosmos-compatible VAE and Text Encoder nodes.

⚖️ License

This model follows the NVIDIA Cosmos License. Please refer to the original model card for full usage restrictions.