deepnet/BestLlamaTest

BestLlamaSN29

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

Merge Details

Merge Method

This model was merged using the passthrough merge method.

Models Merged

The following models were included in the merge:

• coldint/10.5B_v1

Configuration

The following YAML configuration was used to produce this model:

slices:
  - sources:
      - model: coldint/10.5B_v1
        layer_range: [0, 36]
  - sources: # add middle layers with residuals scaled to zero
      - model: coldint/10.5B_v1
        layer_range: [34, 36]
        parameters:
          scale:
            - filter: o_proj
              value: 0.0
            - filter: down_proj
              value: 0.0
            - value: 1.0
              # - sources: # add middle layers with residuals scaled to zero
              #- model: upstage/SOLAR-10.7B-v1.0
              # layer_range: [14, 24]
              # parameters:    
              #scale:
              #- filter: o_proj
              #value: 0.0
              # - filter: down_proj
              # value: 0.0
              # - value: 1.0
  - sources:
      - model: coldint/10.5B_v1
        layer_range: [36, 43]
merge_method: passthrough
dtype: bfloat16