OmniQuant: Omnidirectionally Calibrated Quantization for Large Language Models
Paper • 2308.13137 • Published • 20
3-bit OmniQuant quantized version of Llama-3.3-70B-Instruct for inference with the Private LLM app.
- Downloads last month
- -
Model tree for numen-tech/Llama-3.3-70B-Instruct-w3a16g40sym
Base model
meta-llama/Llama-3.1-70B Finetuned
meta-llama/Llama-3.3-70B-Instruct