AI BriefingTernary-Bonsai-2-27B-gguf: 27B parameters compressed to 2 bits, optimized for local execution
prism-ml/Ternary-Bonsai-2-27B-gguf
About the project
Ternary-Bonsai-2 is a model with 27B parameters quantized to 2 bits. It offers various GGUF formats from F16 to PTQ1_0, enabling immediate execution in local environments such as llama.cpp, Ollama, and LM Studio.
It adopts a hybrid attention structure to improve memory efficiency. With support for CUDA and Metal, inference is possible even on desktops or laptops with limited GPUs. It is specialized for text generation and conversational tasks.
Licensed under Apache 2.0, it allows for free commercial use. Although the total file size is approximately 53GB, 2-bit quantization lowers the hardware barrier to entry compared to existing large models.
prism-ml/Ternary-Bonsai-2-27B-gguf
The original page has no description.
text-generation
This introduction was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report errors, attribution issues, or removal requests via Contact.