Files
Llama-3.2-3B-Instruct-FP8/recipe.yaml
ModelHub XC 06105f88dc 初始化项目,由ModelHub XC社区提供模型
Model: RedHatAI/Llama-3.2-3B-Instruct-FP8
Source: Original Platform
2026-09-27 09:43:13 +08:00

8 lines
144 B
YAML

quant_stage:
quant_modifiers:
QuantizationModifier:
ignore: [lm_head]
scheme: FP8
observer: mse
targets: [Linear]