Capybara: Unified Visual Creation Model

Comprehensive visual generation and editing framework built on diffusion models and transformer architectures. Enables high-quality visual synthesis and manipulation tasks across multiple modalities.

Supported Tasks

• Text-to-Video (T2V)
• Text-to-Image (T2I)
• Instruction-based Video-to-Video (TV2V)
• Instruction-based Image-to-Image (TI2I)

Key Technical Features

• Multi-task support across generation and instruction-based editing
• Distributed inference for efficient multi-GPU processing
• FP8 quantization for memory optimization
• ComfyUI integration with custom nodes
• Automatic prompt enhancement via Qwen3-VL-8B-Instruct

Architecture

• Scheduler configuration
• Text encoders (byt5-small, Glyph-SDXL-v2, LLM components)
• Transformer module (Capybara v0.1)
• VAE for image encoding
• Vision encoder (SIGLIP)

PyTorch 2.6.0 with CUDA 12.6 support. Python 3.11. MIT License.