初始化项目,由ModelHub XC社区提供模型
Model: aisingapore/Qwen-SEA-LION-v4-32B-IT-4BIT Source: Original Platform
This commit is contained in:
27
recipe.yaml
Normal file
27
recipe.yaml
Normal file
@@ -0,0 +1,27 @@
|
||||
quant_stage:
|
||||
quant_modifiers:
|
||||
GPTQModifier:
|
||||
config_groups:
|
||||
group_0:
|
||||
targets: [Linear]
|
||||
weights:
|
||||
num_bits: 4
|
||||
type: int
|
||||
symmetric: false
|
||||
group_size: 64
|
||||
strategy: group
|
||||
block_structure: null
|
||||
dynamic: false
|
||||
actorder: !!python/object/apply:compressed_tensors.quantization.quant_args.ActivationOrdering [
|
||||
static]
|
||||
observer: minmax
|
||||
observer_kwargs: {}
|
||||
input_activations: null
|
||||
output_activations: null
|
||||
format: null
|
||||
targets: [Linear]
|
||||
ignore: [lm_head, 're:.*q_norm', 're:.*k_norm']
|
||||
block_size: 128
|
||||
dampening_frac: 0.01
|
||||
actorder: static
|
||||
offload_hessians: false
|
||||
Reference in New Issue
Block a user