Files
Qwen2.5-14B-Instruct-FP8-W8A8/recipe.yaml
ModelHub XC 25bd811ce1 初始化项目,由ModelHub XC社区提供模型
Model: JongYeop/Qwen2.5-14B-Instruct-FP8-W8A8
Source: Original Platform
2026-07-31 20:23:37 +08:00

33 lines
844 B
YAML

quant_stage:
quant_modifiers:
QuantizationModifier:
config_groups:
group_0:
targets: [Linear]
weights:
num_bits: 8
type: float
symmetric: true
group_size: null
strategy: tensor
block_structure: null
dynamic: false
actorder: null
observer: minmax
observer_kwargs: {}
input_activations:
num_bits: 8
type: float
symmetric: true
group_size: null
strategy: tensor
block_structure: null
dynamic: false
actorder: null
observer: minmax
observer_kwargs: {}
output_activations: null
format: null
targets: [Linear]
ignore: [lm_head]