4bit yml config

#1
by bhuvneshsaini - opened

Why do only this config works in 4bit models but fails with slerp method

models:

  • model: anthonymeo/NEC-merged-best
    parameters:
    density: 0.5
    weight: 0.5
  • model: anthonymeo/llama3.1-merged
    parameters:
    density: 0.5
    weight: 0.5
    merge_method: ties
    base_model: unsloth/Meta-Llama-3.1-8B-Instruct-bnb-4bit
    parameters:
    normalize: false
    int8_mask: true
    dtype: float16

slerp config
slices:

  • sources:
    • model: unsloth/Llama-3.2-3B-Instruct-bnb-4bit
      layer_range:
      • 0
      • 28
    • model: taareshg/Llama-3.2-3B-Instruct-En-Hi-merge-200k
      layer_range:
      • 0
      • 28
        merge_method: slerp
        base_model: unsloth/Llama-3.2-3B-Instruct-bnb-4bit
        parameters:
        t:
    • filter: self_attn
      value:
      • 0
      • 0.5
      • 0.3
      • 0.7
      • 1
    • filter: mlp
      value:
      • 1
      • 0.5
      • 0.7
      • 0.3
      • 0
    • value: 0.5
      dtype: bfloat16

Sign up or log in to comment