OmniQuant: Omnidirectionally Calibrated Quantization for Large Language Models
Paper
•
2308.13137
•
Published
•
18
4-bit OmniQuant quantized version of gemma-2-Ifable-9B for inference with the Private LLM app. This model is currently ranked second on the Creative Writing Benchmark (on January 28th, 2025).
- Downloads last month
- -
Model tree for numen-tech/gemma-2-Ifable-9B-w4a16g128asym
Base model
ifable/gemma-2-Ifable-9B