f3692f2dc2
Model: RedHatAI/Mistral-Small-24B-Instruct-2501-quantized.w4a16 Source: Original Platform
15 lines
435 B
YAML
15 lines
435 B
YAML
DEFAULT_stage:
|
|
DEFAULT_modifiers:
|
|
GPTQModifier:
|
|
sequential_update: true
|
|
dampening_frac: 0.05
|
|
config_groups:
|
|
group_0:
|
|
targets: [Linear]
|
|
weights: {num_bits: 4, type: int, symmetric: true, group_size: 128, strategy: group,
|
|
actorder: null, observer: minmax}
|
|
input_activations: null
|
|
output_activations: null
|
|
targets: [Linear]
|
|
ignore: [lm_head]
|