File size: 228 Bytes
a963cba
 
 
 
 
 
 
952aa3a
1
2
3
4
5
6
7
8
9
default_stage:
  default_modifiers:
    QuantizationModifier:
      targets: [Linear]
      ignore: [lm_head, 're:.*mlp.gate[.].*']
      scheme: FP8
      bypass_divisibility_checks: false
      requires_calibration_data: true