kromcomp/L3.1-Brrv2-8B

brr

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

Merge Details

Merge Method

This model was merged using the Model Stock merge method using nvidia/Llama-3.1-8B-UltraLong-1M-Instruct as a base.

Models Merged

The following models were included in the merge:

• elyn-dev/Llama-3-Soliloquy-8B-v2
• nvidia/Llama-3.1-8B-UltraLong-1M-Instruct + hikikomoriHaven/llama3-8b-hikikomori-v0.4

Configuration

The following YAML configuration was used to produce this model:

base_model: nvidia/Llama-3.1-8B-UltraLong-1M-Instruct
chat_template: llama3
dtype: float32
merge_method: model_stock
modules:
  default:
    slices:
    - sources:
      - layer_range: [0, 32]
        model: elyn-dev/Llama-3-Soliloquy-8B-v2
      - layer_range: [0, 32]
        model: nvidia/Llama-3.1-8B-UltraLong-1M-Instruct+hikikomoriHaven/llama3-8b-hikikomori-v0.4
      - layer_range: [0, 32]
        model: nvidia/Llama-3.1-8B-UltraLong-1M-Instruct
parameters:
  int8_mask: 1.0
tokenizer:
  pad_to_multiple_of: 32