Nexesenex/Llama_3.x_70b_Tess_Dolphin_128K_v1.0

merge

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the Linear DELLA merge method using migtissera/Tess-3-Llama-3.1-70B as a base.

Models Merged

The following models were included in the merge:

• cognitivecomputations/dolphin-2.9.1-llama-3-70b

Configuration

The following YAML configuration was used to produce this model:

merge_method: della_linear
base_model: migtissera/Tess-3-Llama-3.1-70B
models:
  - model: cognitivecomputations/dolphin-2.9.1-llama-3-70b
    parameters:
      weight:
        - filter: q_proj
          value: [0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0]
        - filter: k_proj
          value: [0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0]
        - filter: v_proj
          value: [0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0]
        - filter: o_proj
          value: [0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0]
        - filter: input_layernorm
          value: [0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0]
        - filter: up_proj
          value: [0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0]
        - filter: gate_proj
          value: [0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0]
        - filter: down_proj
          value: [0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0]
        - filter: post_attention_layernorm
          value: [0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0]
        - value: 0
      density: 0.5
      epsilon: 0.1
      lambda: 1.0
  - model: migtissera/Tess-3-Llama-3.1-70B
    parameters:
        weight: 1.0
        density:
          - filter: q_proj
            value: [1, 1, 1, 1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1, 1, 1, 1]
          - filter: k_proj
            value: [1, 1, 1, 1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1, 1, 1, 1]
          - filter: v_proj
            value: [1, 1, 1, 1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1, 1, 1, 1]
          - filter: o_proj
            value: [1, 1, 1, 1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1, 1, 1, 1]
          - filter: input_layernorm
            value: [1, 1, 1, 1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1, 1, 1, 1]
          - filter: up_proj
            value: [1, 1, 1, 1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1, 1, 1, 1]
          - filter: gate_proj
            value: [1, 1, 1, 1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1, 1, 1, 1]
          - filter: down_proj
            value: [1, 1, 1, 1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1, 1, 1, 1]
          - filter: post_attention_layernorm
            value: [1, 1, 1, 1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1, 1, 1, 1]
          - value: 0.5
        epsilon:
          - filter: q_proj
            value: [0, 0, 0, 0, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0, 0, 0, 0]
          - filter: k_proj
            value: [0, 0, 0, 0, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0, 0, 0, 0]
          - filter: v_proj
            value: [0, 0, 0, 0, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0, 0, 0, 0]
          - filter: o_proj
            value: [0, 0, 0, 0, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0, 0, 0, 0]
          - filter: input_layernorm
            value: [0, 0, 0, 0, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0, 0, 0, 0]
          - filter: up_proj
            value: [0, 0, 0, 0, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0, 0, 0, 0]
          - filter: gate_proj
            value: [0, 0, 0, 0, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0, 0, 0, 0]
          - filter: down_proj
            value: [0, 0, 0, 0, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0, 0, 0, 0]
          - filter: post_attention_layernorm
            value: [0, 0, 0, 0, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0, 0, 0, 0]
          - value: 0.1
        lambda: 1.0
dtype: bfloat16
out_dtype: bfloat16
parameters:
  int8_mask: true
  normalize: true
  rescale: true
  filter_wise: false
chat_template: auto
tokenizer:
  source: union