Files
ModelHub XC 6fc1899f0a 初始化项目,由ModelHub XC社区提供模型
Model: neuralmagic/granite-3.1-2b-instruct-quantized.w4a16
Source: Original Platform
2026-09-03 06:44:13 +08:00

12 lines
318 B
YAML

quant_stage:
quant_modifiers:
GPTQModifier:
sequential_update: true
dampening_frac: 0.01
ignore: [lm_head]
config_groups:
group_0:
weights: {num_bits: 4, type: int, symmetric: true, strategy: group, group_size: 64,
observer: mse}
targets: [Linear]