Files
Meta-Llama-3.1-8B-quantized…/recipe.yaml
ModelHub XC 4ce4bb2057 初始化项目,由ModelHub XC社区提供模型
Model: neuralmagic/Meta-Llama-3.1-8B-quantized.w8a8
Source: Original Platform
2026-06-16 11:30:13 +08:00

9 lines
173 B
YAML

quant_stage:
quant_modifiers:
GPTQModifier:
sequential_update: false
dampening_frac: 0.01
ignore: [lm_head]
scheme: W8A8
targets: Linear